It feels strange to finally say it, but with iOS 27, Siri has become more useful than ever. When Apple debuted its voice assistant in the iPhone 4S all those years ago, we had no idea it would become a punchline for so long, paling in comparison to Amazon's Alexa and Google's Home Assistant before being blown away by LLMs like ChatGPT, Claude, and Gemini.

Now, though, the Cupertino firm's AI tech has finally caught up (with a little help from Google's Gemini). Apple Intelligence may have felt like vaporware after its demo a couple of years ago, but now it's the real deal across pretty much all of Apple's 27 operating systems.

Sure, you can ask how many millilitres are in a teaspoon, how many letters are in a word, and all sorts of other relatively generic questions, but the real strength of Siri AI and Apple Intelligence is in its awareness of personal context. Your iPhone knows your contacts, can read your messages and emails, and check your calendar, meaning it can surface the appropriate information when it's needed. Here are some examples.

Before we begin

It's worth noting that you'll need Siri AI on your device. That sounds obvious, but it's not available in all regions, and is notably not coming to the EU anytime soon.

You'll also need to activate it within the Settings menu and ensure it indexes all of the relevant information from your apps. This can take a little while (sometimes more than a few hours), so be patient.

Surface a recommendation

My friends and I have similar music tastes, which means our group chat is always getting fresh links for Spotify. That means I can say "what was that song Phil recommended to me last week?" And have Siri pull a Spotify link from our chat for me to hit play.

It's worth noting that, at the time of writing, WhatsApp hasn't yet been updated for Siri AI, so your mileage may vary, but you can expect it to work with other links shared.

Plan the day ahead

While it is possible to use agentic AI and apps like Slashy to get a quick look at your day, Siri AI can find your appointments, too.

"Show me my meetings for this week" can highlight everything you have coming up, allowing you to easily plan out where you need to be and when.

Just don't forget that you need to enable each Calendar individually in your Settings app, lest you miss an important meeting because your work calendar wasn't ticked on the list!

Pick the best restaurant

Siri's new on-screen awareness means it notes what's going on while you use your iPhone, while Visual Intelligence is ideal for grabbing information from the world around you.

That means you pull up a few restaurants and ask Siri for recommendations based on user reviews, narrow the search to the types of food served, and more.

You can also use Visual Intelligence to instantly translate a menu, or ask for a deeper look at one item in particular. With a quick lookup, you can even find out nutritional information about something you're thinking of ordering.

Find that photo, and expand it

We've all been there. Whether it's a memory from last week, or years ago, we've all had to search for a specific photo in amongst the mess of our Photo Library, and there's a good chance it's not in a folder because who has the time?

Thankfully, asking Apple Intelligence for "all of the photos from the beach in Altea" will surface them. This also ties into the latest editing options, too. Asking Siri AI to "extend this image" will use the Extend tool to fill in around the outside of the image with generative AI, helping reframe your subject, too.

Manage your files

Asking Siri to find a file you've misplaced has saved me more time than I care to admit as a small business owner. Asking Siri AI to find a particular screenshot for a feature or identify the latest version of a file has also been super useful - not just on the iPhone, but also on iPad and Mac.

"Siri, where's that presentation for Tom's Guide?" highlights where it is, and I can even ask it to pull key data from the file itself.

Sure, a lot of this has been possible on Gemini devices for some time, but if you're new to using an AI assistant such as this, there's a ton of ways to learn the ropes. Our advice is to experiment!

More from Tom's Guide

• Nvidia says 'local AI is here' — and it could change how you use AI at home
• OpenAI is paying influencers to convince you to use ChatGPT — here’s what we know
• I tested ChatGPT-6 vs Claude Opus 5.5 with 5 everyday prompts — it wasn’t even close

Apple is slowly rolling out a revamped Health app for iPhone and Apple Watch, and you can try it now by downloading the iOS 27.2 public beta; you’ll also need to run watchOS 27 on your Apple Watch.

One of the coolest new features in the overhauled app is a series of at-home longevity assessments designed to be done in your living room using your iPhone’s camera and Apple Watch’s heart rate sensor.

These assessments provide insights into a user’s strength, balance, flexibility, and more. I tried both the balance and flexibility tests, which take about five to seven minutes each, and gained some valuable, surprising insights into my physical health. Here’s what I learned.

How Apple’s at-home longevity tests work

To perform one of these assessments, users are asked to place their iPhone on the ground, lean up against a wall, and then walk back ten to 12 feet, ensuring their entire body is framed within the front-facing camera’s view.

The assessments use a vision-based AI model to track your body movement as you perform each step. However, no video is actually being recorded during the assessment, and all the AI processing occurs on-device.

What I learned from Apple’s flexibility test

Once your iPhone is set up and you're well-framed in the center of the screen, it’s time to start the assessment.

An on-screen trainer demos the movements for each test with more than enough detail to easily follow along. For example, the flexibility assessment involves five different stretches to thoroughly determine your overall upper and lower body flexibility ratings. Ratings fall into one of four categories: Needs Improvement, Fair, Good, and Excellent.

So, how did I do? My upper body flexibility score was ‘Good,’ but my lower body flexibility score was only ‘Fair,’ lower than I expected. I performed the five tests in the assessment twice, just to be sure I properly followed all the instructions, and both times the results were the same.

Was I aware that my lower body flexibility could use some attention? Honestly, no, not at all. Now that I am, I plan to prioritize more workouts and stretches to improve it.

Helpfully, in addition to a rating, the new Health app also provides fairly comprehensive information on the best ways to make progress in either category, including suggested exercises to try.

My balance is better than my flexibility, according to Apple

The balance assessment is even easier than the flexibility one, involving just one test. Users are asked to stand on one foot, with their other foot suspended in the air behind them for as long as they can without touching the ground, for up to sixty seconds. You do this twice, once for each foot.

Once the countdown begins, calming music starts playing, while a voice periodically provides encouraging words. Balance is scored on the same rubric as flexibility.

Fortunately, this time around I scored an Excellent — again, I did the test twice — which was a lovely little ego boost after my mediocre flexibility performance.

Like an at-home physical exam

These are just two of the 17 longevity assessments found in the new Apple Health app. Once all the assessments are complete, the Health app generates a report based on the results to share with your doctor.

Ultimately, Apple’s new longevity assessment might be foreshadowing what the future of healthcare will look like, where users rely on a mix of AI-powered consumer tech and traditional doctor visits to better understand and improve their well-being.

What are your thoughts on Apple’s new longevity tests in the updated Health app in iOS 27.2? Are you looking forward to trying them yourself? Let me know in the comments below.

More from Tom's Guide

• Apple Watch Ultra 4 first impressions: 5 reasons to buy and 2 reasons to skip Apple’s new rugged model
• Here's how to enable the new Siri AI on your Apple Watch
• Apple just killed one of my favorite Apple Watch features with the launch of watchOS 27

With the Apple Watch Series 12 and Apple Watch Ultra 4, users can now view their total daily step count in real time on the watch face courtesy of Apple’s powerful S11 processor.

While there are a couple of different ways to add the Daily Steps widget to your watch face, below, I’ve explained my favorite way to do it: Using the new Siri Modular watch face in five quick steps.

Don’t have one of the latest Apple Watch models? You can still add a step count widget to the watch face using a third-party option like the Pedometer++ app.

That said, if you do have the Apple Watch Series 12 or Apple Watch Ultra 4, here’s how to unlock Apple’s step count widget in less time than it takes to walk around the block.

How to add the step count widget to your Apple Watch

1. Open the Watch app on your iPhone and tap ‘Face Gallery’ at the bottom, center of the page
2. Select the category ‘New Watch Faces' or search for “Siri Modular’
3. Select your preferred Siri Modular layout and scroll down to ‘Complications’
4. Tap on one of the complications to change it; you’ll find the ‘Daily Steps’ widget toward the top under ‘Activity’
5. With the ‘Daily Steps’ widget added, tap ‘Add to Watch’ to switch to the Siri Modular face

Read on to see full illustrated instructions for each step.

And... that's it! Your personalized Siri Modular watch face with a Daily Steps widget should now appear on your Apple Watch Series 12 or Apple Watch Ultra 4 (the only models that support this feature natively).

Looking for more ways to customize your Apple Watch with the latest watchOS 27 bells and whistles? Here's how to enable the new Siri AI voice assistant on your Apple Watch. Best of all, you don't need one of the latest models – just one that supports watchOS 27.

More from Tom's Guide

• Apple Watch Ultra 4 first impressions: 5 reasons to buy and 2 reasons to skip Apple’s new rugged model
• Apple just killed one of my favorite Apple Watch features with the launch of watchOS 27
• I’ve been writing about the Apple Watch for the past decade: Here’s the rumored update I’m most excited about for the Apple Watch Ultra 4

If you recently updated your Mac to macOS 27 and noticed your available storage shrinking, Apple Intelligence could be taking up more space than you realize.

Mac users on Reddit and Hacker News have reported discovering several gigabytes of AI-related files on their machines, with some users claiming local AI assets are taking up around 10GB or more. That's a significant chunk of storage if you have a 256GB MacBook Air or Mac mini.

There's a reason those files are there. Apple Intelligence doesn't run entirely in the cloud. Some of Apple's AI features rely on models stored and processed directly on your Mac, which is great for privacy but not necessarily great for your available storage.

And if you don't use Apple Intelligence, you might understandably wonder why those files need to be sitting on your SSD at all. Here's what's happening, how to check your Mac and what you can safely do about it.

Why is Apple Intelligence taking up space on my Mac?

One of Apple's big selling points for Apple Intelligence is privacy. Rather than sending everything you do to a remote server, Apple can perform some AI tasks directly on compatible devices. But, that requires downloading AI models to your Mac.

According to MacRumors, once you install macOS 27 Golden Gate, Apple automatically downloads a multi-gigabyte AI model the next time your Mac connects to the internet. In previous versions of macOS, you could opt out. Now, there's no way to stop the download, making the AI models that power Siri and other Apple Intelligence features effectively part of the operating system.

Apple's latest Apple Intelligence architecture includes AFM 3 Core, a 3-billion-parameter on-device model, alongside a more powerful multimodal model called AFM 3 Core Advanced. Other Apple Intelligence features can use Apple's Private Cloud Compute infrastructure when more processing power is required.

The downside of all that on-device processing is that the models themselves need somewhere to live.

macOS manages many of these resources as system assets, which can make it difficult to tell exactly what's using your storage. You may simply notice the amount of space attributed to System Data increasing.

First, check your Mac's storage

Before changing anything, it's worth checking whether you're actually running low on space.

• Click the Apple menu in the top-left corner.
• Select System Settings.
• Choose General.
• Click Storage.
• Give macOS a moment to calculate everything.

You'll see your storage divided into categories such as Applications, Documents, Photos and System Data.

Unfortunately, macOS doesn't provide a neat "Apple Intelligence models: 10GB" entry, so a large System Data category doesn't automatically mean Apple Intelligence is responsible.

Still, take note of how much free storage you currently have. That gives you a baseline before making any changes.

How to turn off Apple Intelligence

If you don't use Apple Intelligence, the safest first step is simply turning it off.

• Open System Settings.
• Select Apple Intelligence & Siri.
• Find Apple Intelligence.
• Toggle it Off.

Confirm that you want to disable Apple Intelligence if prompted. You can also disable Siri AI from this menu if you don't use it.

Don't expect your free storage to jump immediately after doing this. macOS manages system assets automatically, and disabling Apple Intelligence doesn't mean every downloaded model will disappear the moment you flip the switch.

I'd avoid manually deleting anything from macOS system folders just to force the issue.

What about Screen Time?

macOS also includes restrictions for Apple Intelligence and Siri features through Screen Time.

Head to: System Settings > Screen Time > Content & Privacy

Depending on your Mac and settings, you'll find controls governing access to Apple Intelligence and related generative AI features. These restrictions can be useful if you want to prevent someone using the Mac from accessing particular AI features.

However, I wouldn't rely on Screen Time as a guaranteed way to stop macOS from ever downloading an AI-related system asset. Apple doesn't make that relationship particularly clear.

For most people, disabling Apple Intelligence itself is the better place to start.

Don't delete macOS AI files manually

Here's where things get a little complicated. You'll find Terminal commands online that claim to remove Apple's downloaded AI models manually. Some methods even involve booting your Mac into Recovery Mode and deleting files from macOS system directories.

But, I wouldn't recommend doing that. These are system-managed assets, and deleting files from protected macOS directories can cause unintended problems. There's also no guarantee macOS won't download an asset again if the operating system decides it's required.

Saving a few extra gigabytes isn't worth potentially breaking something in macOS.

If you're critically low on storage, you're better off using Apple's built-in storage management tools first.

Still missing storage? Check Time Machine

Apple Intelligence might not be the only reason your available storage seems lower than expected. If you use Time Machine, macOS can also keep local snapshots on your Mac. These allow you to recover files even when your Time Machine backup drive isn't connected.

Apple normally manages this storage automatically and considers snapshot space available for use when necessary.

That means the amount of storage macOS reports can occasionally be confusing, particularly if you've recently deleted a large amount of data and expected your free-space number to immediately increase.

You can see whether local snapshots exist by opening Disk Utility, selecting your Mac's volume and looking for APFS snapshots.

Again, I wouldn't start deleting snapshots unless you're actually experiencing a storage problem.

Bottom line: Should you turn off Apple Intelligence to save space?

If you regularly use Writing Tools, Siri's newer AI features or other Apple Intelligence capabilities, the storage they consume is essentially part of the cost of having those features available locally.

But it's totally different if you never use them. That's especially true for anyone with a 256GB Mac. Losing even 10GB represents roughly 4% of your entire advertised storage capacity — before accounting for macOS itself, applications, photos, videos and everything else you keep on your computer.

Until Apple gives users more granular control over which AI models are stored locally, the safest option if you don't want Apple Intelligence is to simply turn it off, keep an eye on your storage and let macOS manage the underlying system assets rather than trying to rip them out yourself.

More from Tom's Guide

• Nvidia says 'local AI is here' — and it could change how you use AI at home
• I replaced ChatGPT with Siri AI for a day — here's what surprised me
• AI could be costing you money: new study finds chatbots get most financial questions wrong

Apple settled a $250 million class action lawsuit over claims of false advertising about Siri's AI capabilities in May of this year. The settlement allows up to 36 million iPhone users in the U.S. to receive up to $95.

That window is finally opening, but it runs for only three months, starting Monday, September 21. Here's what you need to know to claim your share of the settlement.

How we got here

The lawsuit was initially filed in California in 2024 by Peter Landsheft after Apple promoted its more "personalized Siri" and the introduction of Apple Intelligence. The National Advertising Division bolstered the lawsuit, concluding that Apple misled users by claiming the new Siri was "now available."

Apple filed the settlement, which includes no admissions of wrongdoing, on May 5. "We resolved this matter to stay focused on doing what we do ⁠best, delivering ​the most innovative products and ​services to our users," the company said in a statement.

“False advertising settlements aim to raise the floor of truth-telling, protect fair competition, and increase consumer choice. The $250 million settlement fund represents the largest false advertising settlement in history, and we are proud to have achieved this result for Apple customers," Ryan Clarkson, one of the lawyers on the case, said in a statement to Tom's Guide.

Eligibility and how to claim

The settlement covers roughly 36 million iPhones, including the iPhone 16, iPhone 15 Pro and iPhone 15 Pro Max. All of which must have been purchased in the U.S. between June 10, 2024, and March 29, 2025.

The window for entering a claim starts September 21 and closes on December 21, 2026.

You can enter a claim by going to SmartphoneAISettlement.com. You will need your device's serial number. You can find the number by opening Settings, going to General, tapping About, and scrolling to Serial Number.

As for payout, claimants will receive at least $25 per device and up to $95, depending on how many claims are submitted.

Final approval by a judge won't happen until February 24, 2027, and Apple will have a set number of days after that to issue payments. Typically, those windows are 180 days.

More from Tom's Guide

• Apple increased the price of iPhone 18 Pro battery repairs — here's how much you'll pay now
• Siri in iOS 27 actually lives up to the hype — here are 3 truly unique uses
• Apple Watch Ultra 4 review: I ran 22 miles to test Apple’s ultimate adventure watch

Apple elected to settle a $250 million class action lawsuit back in May that means up to 36 million iPhone owners are eligible for a payout of up to $95 per device. The lawsuit alleged Apple misled customers with false advertising over the arrival of Siri AI and Apple Intelligence.

Apple denied any wrongdoing, but did decide to settle the lawsuit to resolve the claims swiftly. In a statement, it said this meant it could "stay focused" on "delivering the most innovative products and services to our users."

The settlement allowed each eligible person who filed a claim to receive a per-device payment of $25, which could increase up to $95 depending on the claim volume.

More from Tom's Guide

• I've been testing the iPhone 18 Pro Max and iPhone 18 Pro — your 10 biggest questions answered
• Apple added 5 new lock screen tools in iOS 27 — here's what to change first
• I spent 30 minutes with the iPhone Duo — 3 things you just don’t get from the pictures

Over the past few years I have spent an inordinate amount of my time sitting in virtual queues, waiting for my chance to buy something. Recently my phone was sitting in one for around four hours as I attempted to book a family trip to Disneyland Paris. Since it was the day bookings opened, it seems as though half of Europe was doing the same thing.

Waiting is bad enough, but it was made worse by the fact I got distracted partway through waiting and missed my window. That means starting the whole process again, and mentally kicking myself for letting myself get distracted attempting to troubleshoot my temperamental 3D printer. This is far from an isolated case for me, either.

So you can imagine my excitement when I found out about the new "Notify Me" feature in iOS 27, with the promise that it can monitor webpages and alert you to any changes. Unfortunately spending time with Notify Me shows that Apple hasn't really gone far enough.

What Notify Me is, and how it works

Notify Me is an artificial intelligence feature for Apple's Safari browser. The idea is that you can single out a webpage and have Apple's AI tell you when things change. That can be general changes, or it could be something specific that you describe with a written prompt.

The main thing to remember about Notify Me is that the AI doesn't monitor that page all the time. Instead you have to tell it when it should come back and see what's changed. Your options are to do this hourly, daily, weekly or monthly. You also get to set the exact time, day or date (depending on the frequency) that the AI goes back to check.

It's a very simple feature that can be accessed from the Safari menu bar which sits on the left-hand side of the URL bar — which is typically at the bottom of the screen on iPhone and iPad. It's easy to find and use, which is more than I can say for a lot of features in iOS 27.

From a practical stand point, Notify Me has everything going for it. Apart from the fact its usefulness is currently rather limited.

Hourly checks aren't enough to be truly useful

I'm not saying that Notify Me is useless in its current form, quite the opposite in fact. The tool is ideal for keeping tabs on certain pages in the long-term. Use cases include keeping tabs on restocks, or checking for discounts on specific products. The pages that you would only look at periodically, checking in very briefly before going back to whatever you would normally be doing.

Notify Me saves you the effort of remembering to do this yourself, especially if you want to check on some sort of schedule. You don't need to take time out of your day to look, when you have an AI that has been instructed to take care of that for you.

But in the short term? Notify Me's scope is still a little too limited for my liking.

Typically getting to the front of a virtual queue means you get some kind of grace period to complete whatever transaction you're doing. In the case of my trip it was 20 minutes, but I've seen them vary anywhere between 10 minutes and an hour. Obviously, having a tool that refreshes a page once an hour at most is not ideal for dealing with these kinds of situations. By the time the AI checks the page, it'll already be too late and you'll have missed your spot.

Really it would be better if Apple expanded Notify Me for much smaller periods of time. Five to ten minutes would expand its usefulness by a lot, since it would be as close to real-time monitoring as we can hope for with a free AI tool. But even increasing page refreshes to 30 or 15 minutes would make it a more useful and versatile tool.

Bottom line

Notify Me is definitely a useful tool, but its scope is too limited right now. The fact that you can instruct AI to check webpages for any kind of change you can describe is pretty incredible. But it's very much a long-term tool, that a lot of people probably won't get much use out of. It certainly has the potential to offer a lot more, though.

The question is whether Apple can actually do that. AI needs a lot of resources to run, and the more you utilize it the more pressure it's going to put on your hardware. In fact, we don't know where this processing is done or how it affects your phone, so there's no telling what sort of stress checking a page multiple times an hour might inflict. Even if Apple has just released the fastest phones we've ever tested.

But I hope that this is just the start of what Apple has planned, and that Notify Me will be getting an expansion in future versions of iOS and Apple Intelligence. After all, the less we all have to sit in virtual queues to try and get tickets to the next big concert or event, the better.

More from Tom's Guide

• I tested the iPhone 18 Pro Max's variable aperture camera — here's the results
• Apple increased the price of iPhone 18 Pro battery repairs — here's how much you'll pay now
• Siri in iOS 27 actually lives up to the hype — here are 3 truly unique uses

I’ll be the first to admit I stopped relying on Siri years ago for anything more than quick kitchen timers or checking the weather. Like many iPhone users, I had grown accustomed to its limitations.

But after spending time testing the new Siri experience in iOS 27, I'm actually shocked by how much it's improved. Instead of just answering basic questions, Siri finally feels full integrated into your phone.

If you've been ignoring the feature like I was, here are the three unique tools that actually changed my daily routine.

More from Tom's Guide

• iOS 27: 7 features you need to try first after updating
• I tried the new iOS 27 Safari features — here's 3 settings worth using
• iOS 27 Photos features that transform your pictures — here's how to use them

At last, an AI-powered version of Apple's Siri voice assistant is available for the Apple Watch thanks to the public release of watchOS 27, the newest Apple Watch software.

In my experience so far, Siri AI is more natural in its tone, making conversing less awkward. More importantly, the feature is much better at answering complex, context-based questions with a reasonable degree of reliability.

Here's how to load Siri AI on your Apple Watch and add a dedicated Siri AI widget to your watch face in seven easy steps. Plus, a list of eligible models.

How to enable Siri AI on your Apple Watch

1. Make sure your Apple Watch and iPhone are compatible with Siri AI
2. Update your phone to iOS 27; update your Apple Watch to watchOS 27
3. Turn on Siri in your phone’s settings and join the Siri beta waitlist
4. Turn on Siri in your Apple Watch’s settings
5. Open the Watch app on your iPhone, tap Display and Brightness > Face Gallery
6. Tap “New Watch Faces” and/or search for the “Siri Modular” face
7. Customize the Siri Modular face to your liking and tap “Add to Watch” when finished

Read on to see full illustrated instructions for each step.

And there you have it. Siri AI is now enabled on your Apple Watch, and the Siri AI assistant app is now fixed on your watch face for easy access. Of course, you can also open Siri by double-tapping the Digital Crown or using a gesture control like wrist raise.

Siri AI is still in beta, so your experience with the AI assistant might not be as smooth as the final product. Moreover, getting off the Siri AI beta waitlist can take anywhere from a few minutes or hours to several days. So, be patient.

More from Tom's Guide

• Apple Watch Ultra 4 first impressions: 5 reasons to buy and 2 reasons to skip Apple’s new rugged model
• We asked 368 people what they want from the Apple Watch Ultra 4. Here’s the resounding answer (and it’s not AI)
• I’ve been writing about the Apple Watch for the past decade: Here’s the rumored update I’m most excited about for the Apple Watch Ultra 4

iPadOS 27 is here for the best iPads that support it! Like iOS 27 and macOS 27, iPadOS 27 focuses on overall system refinements and deeper AI integration. That last bit is especially noteworthy, as there’s now a dedicated Siri AI app that functions like a standard chatbot. Beyond answering questions, Siri AI can also take actions on your behalf since it’s integrated into the operating system and works with Apple’s applications.

I’ve been using iPadOS 27 since the beta launched over the summer, and have generally had a smooth experience with the latest OS update. Though I’m not a big AI user, the updated Siri can be handy — especially as it can do things on your behalf. And if you have kids, I’m sure you’ll appreciate the new child safety features.

iPadOS 27 isn’t a game-changer, but features like the overhauled Siri AI, the Liquid-Glass slider, and other quality-of-life improvements make using Apple’s tablets feel smoother and more personalized. Here is my full review of iPadOS 27.

iPadOS 27 review: Compatible iPads

Here are all the iPads that support iPadOS 27.

• iPad Pro (M4 and later)
• iPad Pro 12.9‑inch (4th generation and later)
• iPad Pro 11‑inch (2nd generation and later)
• iPad Air 13-inch (M2 and later)
• iPad Air 11-inch (M2, M3, and M4)
• iPad Air 11-inch (4th generation and later)
• iPad (A16)
• iPad (9th generation and later)
• iPad mini (A17 Pro)
• iPad mini (6th generation)

iPadOS 27 review: Biggest features

The new Siri

Apple Intelligence is finally ready for prime time on the company’s devices. Thanks to the new Siri AI app, you can now interact with Apple’s virtual assistant the same way you would with a chatbot like ChatGPT or Google Gemini. Apple Intelligence runs on both servers and your device to offer better image understanding and more.

There is now a dedicated Siri AI app, which looks and functions like a regular chatbot app. In this window, you can ask Siri questions or have it perform tasks for you. For example, I used Siri to send a message to one of my iMessage group chats, and it did so after I approved it. Sure, I could have just done this directly in the group chat, but it’s a good demonstration of what the new Siri can do.

Siri AI also understands what's on your screen. When I had the Tom’s Guide homepage open on Google Chrome, it correctly detailed what was on the screen, including all the articles. It couldn’t open the articles or read them aloud (despite saying it could), but since I’m testing a beta version, I’ll give it an A for effort, even if it failed to complete the task.

Siri is most useful when you’re using Apple’s own apps and programs. It can create calendar events, set reminders, play music, and more. For example, when you’re looking at a schedule, you’ll see an “add to calendar” button pop up. When you right-click certain folders, files, or apps, you’ll see an “Ask Siri” option at the bottom. Selecting this will open a window where you can start texting or talking with Siri.

Though Siri can’t yet interact with many third-party applications (such as those from Google), it’s a huge improvement over the old Siri.

Liquid Glass improvements

Liquid Glass hasn’t gone anywhere, as Apple is leaning more into the UI with its latest operating system updates. However, Apple is now giving you more control over how it looks with a new slider that adjusts opacity. You can’t eliminate the Liquid Glass look, but you can minimize it significantly.

Other improvements include more diffuse, complex background visuals, so text stands out more. You’ll also notice that the edges of icons are more pronounced to help further separate them from the background. The uniform corner radius of all app windows also helps them stand out more.

Even if Liquid Glass hasn’t gone anywhere, even detractors should find the updates tolerable since they do greatly improve overall readability and visibility. The Liquid Glass slider is a nice touch.

I've never had a problem with Liquid Glass because I think it’s an interesting visual effect. That said, I’m glad that folks can now fine-tune it to their liking. The slider is especially neat because it’s easy to use and clearly shows how much of a difference there is when you go from one extreme to the other.

Child safety features

If you’re a parent, you’ll be happy to know that Apple is expanding its parental control features with iPadOS 27. This way, you can let the kiddo use Apple’s tablet without worrying too much about the kind of content they’re watching and for how long they’re using the slate.

These updates include a redesigned, more intuitive Screen Time Dashboard with one-tap controls and expanded safeguards. There are also Time Allowances for app categories, Ask to Browse website approval, Communication Safety that blurs gore or violence, and one-tap access management. You can tailor these features to suit your child.

I don’t have kids, but I did find the child safety features simple enough to use and customize. Given how much kids love iPads, it’s smart for Apple to keep refining these safety features.

iPadOS 27 review: Other features

• Spatial Reframing: As the name suggests, you can now reframe a photo after it’s been taken thanks to this feature. You can also expand your shots with the Extend Tool and remove objects with the Clean Up tool.
• Upgraded Image Playground: Image Playground can now generate higher-quality images. You can also modify and transform existing images by describing what you want.
• Siri in Notes: Siri AI can assist with your handwritten or typed notes in the Notes app by (for example) organizing bullet points into a formatted agenda or turning handwritten lectures into a detailed study guide.
• Safari updates: In Safari, tabs can now be automatically grouped by topic, making related pages easier to find. With Safari Notify Me, you can receive an alert if there’s a price change or a restock for an item you’ve been eyeing.
• Describe Shortcuts: Instead of manually building complex workflows, you can now describe what you need, and Shortcuts connects actions across multiple apps to make it happen. This is useful if you want to set specific routines, such as launching a specific app when you connect a peripheral.

iPadOS 27 review: Verdict

I use an iPad Pro every single day and have largely been pleased with iPadOS 27. Again, it won’t set the world on fire with revolutionary new features, but the upgraded Siri AI and deeper Apple Intelligence integrations help make iPads that support the update more useful. This is especially true if you use Apple’s own apps often.

As I said in my macOS Golden Gate review, iPadOS 27 feels like a holding-pattern update for whatever big OS refresh the folks in Cupertino are cooking. But there’s also nothing wrong with not fixing what isn’t broken, so it’s fine that this isn’t a huge update. If you own a compatible iPad, you should update as soon as you can to test the new features yourself!

Apple's big week of operating system updates is finally here with the new iPadOS 27. This year's update isn't groundbreaking; instead, it focuses on Apple Intelligence, which should make for a more personal, customized experience.

As with iOS 27 and macOS Golden Gate, the main attraction is the reinvented Siri, which finally lives up to its potential. Siri AI, as it’s now called, is context-aware and understands what’s on your screen. Because it’s built into the OS, it can directly control and manage apps. It also adds new child safety features, improved readability for Liquid Glass, new photo-editing tools, and more.

Here’s how to download iPadOS 27.

How to download iPadOS 27

As always, before you install a major update like iPadOS 27, back up your iPad in case anything goes wrong and you need to return to iPadOS 26.

Not every iPad will be compatible with the new operating system, but if your tablet is, the upgrade is free.

Here is the full list of iPads that are compatible with iPadOS 27:

• iPad Pro (M4 and later)
• iPad Pro 12.9‑inch (4th generation and later)
• iPad Pro 11‑inch (2nd generation and later)
• iPad Air 13-inch (M2 and later)
• iPad Air 11-inch (M2, M3, and M4)
• iPad Air 11-inch (4th generation and later)
• iPad (A16)
• iPad (9th generation and later)
• iPad mini (A17 Pro)
• iPad mini (6th generation)

More from Tom's Guide

• I just downloaded macOS 27 public beta — skip Siri AI and try these 5 hidden features instead
• Haptic vs. Mechanical trackpads: I put them both through a gauntlet, and one almost made me throw my laptop into the sun
• Exclusive: I asked Lenovo’s GM to predict the future of monitors, and it's the end of our chaotic, cable-cluttered desks and monitors that fry your retinas

It's mid-September, which means Apple is now releasing its 2026 operating systems, including the full version of macOS Golden Gate.

If you're interested in Apple Intelligence and the new Siri AI, you should try Golden Gate. There aren't many new features, but it improves Liquid Glass readability, offers easier customization, and delivers smoother performance for a better overall experience on the best MacBooks.

The AI features, marking the end of Apple's Intel era, are powered by Google Gemini and designed to work within Apple's ecosystem.

To make the update process as painless as possible, I've put together this quick, easy guide on how to download macOS Tahoe, with photos to walk you through it.

Supported Macs

As with most macOS releases, Golden Gate is a free upgrade for all supported Macs. Not every device that ran macOS Tahoe can run Golden Gate.

Here's the full list of Macs that can install and run macOS Golden Gate:

• MacBook Neo (2026)
• MacBook Air with Apple silicon (2020 and later)
• MacBook Pro with Apple silicon (2020 and later)
• iMac with Apple silicon (2021 and later)
• Mac mini with Apple silicon (2020 and later)
• Mac Studio with Apple silicon (2022 and later)
• Mac Pro with Apple silicon (2023 and later)

Important before you update: back up your Mac! Apple has the built-in Time Machine feature, or you could use a reputable third-party service like Backblaze — our guide to the best cloud storage solutions offers even more recommendations for you.

How to download macOS Golden Gate

Apple makes downloading its macOS updates fairly straightforward.

That’s all that’s required to download macOS Golden Gate.

Your Mac will restart once the download finishes. After that, you can start trying out all the new features. Depending on your connection speed, the download might take a few minutes, so be patient.

More from Tom's Guide

• 7 Apple products expected to be announced in October — touchscreen MacBook Ultra, HomePad and more
• iOS 27: 7 features you need to try first after updating
• iPhone Duo hands-on: Apple nailed it — and just put everyone else on notice

iOS 27 is now rolling out to compatible iPhones, and a quick glance at the official list tells you that there are a lot of them this year. The thing is, getting the most out of iOS 27 is a lot more than just being able to run the software, and the things you get from iOS 27 will all depend on what kind of hardware your iPhone has on offer.

iOS 27 itself works on devices as old as the iPhone 11, which was first released in 2019. The software has the same requirements as iOS 26, so if your phone runs that you should be able to install the latest iOS update without issue. Unfortunately, some of those older devices have to go without Apple Intelligence, while AI-friendly iPhones still won't gain access to everything the new Siri AI has to offer.

Here's a rundown on the requirements you need to run iOS 27 and all the different tiers of AI Apple now has to offer.

iOS 27

iOS 27 is available on a crazy 33 different iPhone models, dating as far back as the iPhone 11 and coming all the way up to the newly announced iPhone Duo, iPhone 18 Pro and iPhone 18 Pro Max.

Better still, Apple is delivering the update to 30 of those devices today (September 14), letting you upgrade immediately. There are none of the delays of slow-burning rollouts that you'd normally get with Android phones. The only three exceptions are the newest iPhone models, which come running iOS 27 right out of the box.

The list of iOS 27 compatible phones is as follows:

• iPhone 18 Pro
• iPhone 18 Pro Max
• iPhone Duo
• iPhone 17
• iPhone Air
• iPhone 17 Pro
• iPhone 17 Pro Max
• iPhone 16
• iPhone 16 Plus
• iPhone 16 Pro
• iPhone 16 Pro Max
• iPhone 16e
• iPhone 15
• iPhone 15 Plus
• iPhone 15 Pro
• iPhone 15 Pro Max
• iPhone 14
• iPhone 14 Plus
• iPhone 14 Pro
• iPhone 14 Pro Max
• iPhone 13
• iPhone 13 mini
• iPhone 13 Pro
• iPhone 13 Pro Max
• iPhone 12
• iPhone 12 mini
• iPhone 12 Pro
• iPhone 12 Pro Max
• iPhone 11
• iPhone 11 Pro
• iPhone 11 Pro Max
• iPhone SE (both 2020 and 2022 models)

Apple Intelligence

Those older iPhones may be able to support iOS 27, but that doesn't mean they are able to use Apple's AI suite — also known as Apple Intelligence. Apple likes to push all AI processing on devices where possible, and that means strict memory requirements. This means any phone that doesn't have at least 8GB of RAM can not run Apple's AI.

So far 15 phones can run Apple Intelligence, starting with iPhone 15 Pro:

• iPhone 15 Pro
• iPhone 15 Pro Max
• iPhone 16
• iPhone 16 Plus
• iPhone 16 Pro
• iPhone 16 Pro Max
• iPhone 16e
• iPhone 17
• iPhone 17 Pro
• iPhone 17 Pro Max
• iPhone Air
• iPhone 17e
• iPhone 18 Pro
• iPhone 18 Pro Max
• iPhone Duo

All of those phones are capable of running Siri AI, as are iPads and MacBooks with an Apple M-series chip. However they aren't all able to access everything Siri AI has to offer.

Siri AI

The bulk of Siri AI's features are standard Apple Intelligence features. That means they're available to all iPhones with at least 8GB of RAM. However two of the promised Siri AI features run on Apple's "most powerful on-device model", known as AFM Core Advanced. This model needs at least 12GB of RAM to operate.

The specific features are Siri's Expressive Voices and the advanced dictation system. Currently only six iPhones have the hardware necessary to power them:

• iPhone 17 Pro
• iPhone 17 Pro Max
• iPhone 17 Air
• iPhone 18 Pro
• iPhone 18 Pro Max
• iPhone Duo

Other Apple devices that run the AFM Core Advanced include iPad Pro (2024 or later), iPad Air (2026 or later), MacBook Pro (2023 or later) and MacBook Air (2024 or later).

Siri AI isn't available on iPhones and iPads in the EU and China, over regulatory hurdles. While Apple hasn't spoken about the delay in a Chinese launch, it has made it repeatedly claimed that the European Union's Digital Markets Act is holding upSiri AI's launch in the region — evidently due to regulators not accepting "any of Apple’s proposed solutions to bring Siri AI to the EU while safely supporting other virtual assistants."

Like many Apple spats with the EU, it all comes down to the topic of safety and privacy. Apple believes having its rules in place promotes a safer environment, while the European Commission has reportedly disputed those claims.

More from Tom's Guide

• iPhone Duo vs Galaxy Z Fold 8 — 7 ways Apple beats Samsung's foldable
• iPhone Duo could be followed by iPhone Duo Max and iPhone Duo SE — here's what we know
• iPhone 18 Pro and Pro Max variable aperture camera explained — here's what it actually does

Among the big new features Apple revealed with the Apple Watch Series 12 and Apple Watch Ultra 4 were a few that pricked the ears of those with privacy concerns. Apple's Audio Intelligence is a suite of AI-powered features that use the Apple Watch to listen in on what's happening around you, so that Siri can provide more proactive advice, as well as help you remember what it was you just heard.

The most controversial of these features were Siri Recap, which allows your watch to listen in on conversations you have throughout the day, and uses AI to take notes and generate summaries, and Live Rewind, which lets you go back 15 seconds and view a transcript of what was just said.

The idea behind this is that Apple can "jog your memory" on events from your day, and make sure you can stay present in the moment. Naturally there are a lot of people who aren't happy about the idea that someone's Apple Watch could be listening in on conversations without warning.

Apple has anticipated this, and when we spoke to John Ternus, the new Apple CEO, and Greg Joswiak, Apple's senior vice president of worldwide marketing, they revealed their approach to Recap's privacy at length.

"Whenever we do anything, we always do it with an eye towards privacy and security," Ternus said. "It's something we take very seriously and approach very thoughtfully.

On the subject of Siri Recap, Ternus pointed out that it's been possible to record audio and create transcripts of what people have said for a while — both on smartwatches and phones.

Ternus makes it clear that recap does not produce a recording or a transcript, with Jozwiak adding that Apple can't access what doesn't exist. "If we don't have a recording, it means no one else has the recording. It can't get hacked. It can't get delivered to somebody."

It sounds unrealistic, but Jozwiak does confirm that Apple had to develop a whole new chip to make Recap possible in this manner. In the process, he sees it as a feature that allows people to concentrate on what they're doing, while being able to generate "higher level notes" that aren't indiscriminately documenting whatever sensitive information is being said.

"We wanted to do this in a way that people feel comfortable using it, said Ternus. "And people around the user feel comfortable about its existence."

Siri Recap's privacy features

Apple has also posted a sizable explainer discussing all the different privacy features Recap incorporates.

The big privacy feature here is that any audio that the Watch hears is sent to Apple's Secure Enclave on the watch itself or your paired iPhone. Apple claims that it can't be accessed by the user, Apple, apps or even the operating system itself.

The audio also isn't recorded or stored, though the processed data is summarized and formatted by Apple's Private Cloud Compute. Apparently this involves condensed versions of the original conversation that are less than half their length, reducing the contents to the core, rather than a full-length transcript.

While the Private Cloud Compute data is not accessible by Apple or anyone else, it does still mean your data is being processed on the cloud rather than just on your device. Which may be too risky for some of you to agree with.

Other features include a lack of speaker attribution, which means nobody will be directly quoted or attributed to. Siri is also going to censor what it hears, and will automatically omit anything it deems harmful or sensitive. Apple says this includes language promoting self-harm, hate speech, financial data, authentication and personally identifying data and "government-assigned identifiers." So it won't be making a record of your social security number.

Summaries are automatically deleted after seven days unless you save them to the Siri app. Like all Siri data, those saved summaries can then be deleted from all synced devices at any time — though you can also export them to other apps, including Apple Notes. Syncing is also encrypted and done via iCloud, meaning only your own trusted devices can access those summaries.

Recaps aren't running all the time, and you do need to opt in to actually set them up. How they record depends on you, with Apple promising you can set up a schedule or manually toggle it on and off in the Control Center.

Because it's an AI feature, Apple has the usual caveat explaining that the summaries aren't going to be 100% accurate all the time. That doesn't fill me with a whole lot of confidence, especially after the fiasco with Apple News summaries that had to be recalled from iOS.

Live Rewind

If you've ever been pulled in two directions at once, you've inevitably had to ask "what did you just say?" to someone. Live Rewind is meant to help jog your short-short-term memory, by displaying a text snippet of the last 15 seconds of conversation when you double-press the Digital Crown on your Apple Watch.

To address privacy concerns around this feature, Apple says that these text snippets "disappear permanently from your Apple Watch approximately 30 seconds after the Apple Watch display dims," unless you purposely save them.

This doesn't fix everything

Apple suggests that people "be mindful" of everyone around them, especially if there are sensitive conversations happening. Though that may be giving people too much credit.

There are plenty of people out there that are considerate of other people's thoughts and opinions, but there are also plenty who aren't. They either forget recap is running, or straight up do not care that other people may not like it.

Admittedly, both Ternus and Jozwiak have a point: there are features that offer this sort of thing without any emphasis on privacy. Then again, it is more likely that people take issue with the implied always-listening approach Recap takes. I personally feel it would make a lot more sense to restrict it to manual control.

Whichever way you look at it, though, I guess it's still a better option than seeing people pull out their phones to video every little unusual thing they come across.

More from Tom's Guide

• We've tested every Apple Watch and these are the 3 models I recommend right now
• Apple Watch Series 12 vs Apple Watch Ultra 4: Which new Apple Watch should you buy?
• All the new audio features on Apple Watch explained

Apple just announced a new Mac Studio built around the M5 Max and an all-new M5 Ultra. It’s still the same compact desktop that sits on your desk, but this time it focuses on local AI performance and the heaviest pro workloads.

The company is claiming up to 4.3x faster AI performance, up to 1.8x faster graphics, up to 1.3x faster CPU speeds, higher memory bandwidth, and up to twice the storage speed compared with the previous generation. You also get Wi-Fi 7 and Bluetooth 6 for the first time on Mac Studio, plus Thunderbolt 5 ports. It pairs with Studio Display and Studio Display XDR and runs the upcoming macOS 27 Golden Gate with the next version of Apple Intelligence, including Siri AI.

Pre-orders start today, with units scheduled to start arriving September 22. Here’s everything you need to know about the new Mac Studio.

Mac Studio: Price & availability

The Mac Studio with M5 Max starts at $2,499 (or $2,299 for education), while the Mac Studio with M5 Ultra starts at $5,499 (or $5,099 for education). Apple offers configure-to-order options on its site. The top 512GB unified memory configuration is expected in late October.

Pre-orders begin today on Apple’s website and in the Apple Store app in 30 countries and regions, including the U.S. Units begin shipping on September 22.

Mac Studio: Design

The new Mac Studio keeps the familiar compact form factor that fits on a desk. Apple didn’t announce major exterior changes, so it should look and feel like the previous model.

What is new is the connectivity. Thunderbolt 5 ports deliver up to 120Gb/s, which is useful for fast external storage, PCIe expansion chassis, and high-performance hubs. You can also cluster multiple Mac Studio systems together over Thunderbolt 5 with RDMA support. Apple says a cluster of four can deliver up to 3x faster distributed AI inference than a single machine.

Wireless gets an upgrade too. The Apple N1 chip brings Wi-Fi 7 and Bluetooth 6 to the Mac Studio for the first time. There’s also genlock support over USB-C for tighter synchronization between a display and a camera (Apple specifically calls out the iPhone 17 Pro).

On the display side, the Mac Studio can drive up to eight displays, or up to four Studio Display XDR units at full 5K resolution and 120Hz.

Storage is faster as well. Apple switched to a next-generation SSD architecture based on PCIe Gen 6 and claims up to 2x the performance of the previous Mac Studio for loading projects, transferring files, and working with large language models.

Mac Studio: Specs & performance

There are two chips here. M5 Max is the more affordable option for serious pro work. M5 Ultra is the maxed-out version for the most extreme workloads.

Apple says the GPU is up to 50% faster than the previous generation. It also includes third-gen hardware-accelerated ray tracing and enhanced shader cores. The Media Engine handles hardware-accelerated H.264, HEVC, ProRes, and AV1 decoding, which should help with uncompressed 8K footage and multiple concurrent video streams.

Other claims for M5 Max include:

• Up to 10.7x faster LLM prompt processing in LM Studio versus Mac Studio with M1 Max (3.9x versus M4 Max)
• Up to 7.4x faster text-to-image performance versus M1 Max (3.5x versus M4 Max)
• Up to 5.3x faster Magic Mask in Blackmagic Design DaVinci Resolve Studio versus M1 Max (3x versus M4 Max)
• Up to 3.5x faster DNA sequencing basecalling in Oxford Nanopore MinKNOW versus M1 Max (1.9x versus M4 Max)

Apple positions the Ultra as the chip with no real equivalent. Multithreaded CPU performance is listed at up to 1.3x higher than the M3 Ultra. Graphics are up to 1.8x faster than the prior generation.

Neural Accelerators appear on the Ultra for the first time, helping deliver the peak AI numbers. The Media Engine has twice as many video encode and decode blocks as the M5 Max, enough to play up to 33 streams of 8K ProRes 422 at 30 fps simultaneously.

Other claims for M5 Ultra include:

• Up to 15.4x faster CopyCat ML training in Foundry Nuke versus Mac Studio with M1 Ultra (3.3x versus M3 Ultra)
• Up to 9.8x faster LLM prompt processing in LM Studio versus M1 Ultra (4x versus M3 Ultra)
• Up to 8.2x faster text-to-image performance versus M1 Ultra (4.3x versus M3 Ultra)
• Up to 4.7x faster scene rendering in Maxon Redshift versus M1 Ultra (1.7x versus M3 Ultra)

On the AI side, Apple highlights Neural Accelerators in every GPU core to speed up matrix multiplication. M5 Max is rated up to 3.9x faster AI performance than the prior generation. M5 Ultra reaches up to 4.3x the peak AI compute of M3 Ultra and 9.8x that of M1 Ultra. Combined with the large unified memory pools, the pitch is that you can run substantial open-weight models entirely on-device.

Apple is also pushing two frameworks: Core AI (a new architecture optimized for unified memory, CPU, GPU, and Neural Engine) and the existing open-source MLX framework. Both are meant to make local training, fine-tuning, and deployment more practical on Apple silicon.

Mac Studio: Outlook

The new Mac Studio is for people who want to run large AI models, complex video timelines, heavy 3D/VFX work, or big data jobs on a single compact desktop instead of relying on the cloud. The jump in memory (especially the 512GB Ultra configs), the faster storage, Thunderbolt 5 clustering, and the Neural Accelerator hardware are the main selling points.

Whether the real-world gains match Apple's claims will depend on the specific apps and workloads you actually use. The base prices are in line with previous high-end Mac Studios, though they're certainly not cheap. We’ll see what the Mac Studio can do when we get it into our testing lab in September, so stay tuned for more!

Apple just announced a new Mac mini that’s available with either the all-new M6 chip or the more powerful M5 Pro chip. You’re still getting what looks to be one of best mini PCs and the most affordable Apple desktop, but with this new generation of Mac mini, overall speed and the ability to run AI models locally seem to be the main focus.

Given how popular the Mac mini M4 became with AI enthusiasts, the company has given the base model a major boost in that department. According to Apple, the Mac mini with M6 delivers up to 4x faster AI performance, 2x faster storage and graphics as well as 40% faster CPU performance. When it comes to the Mac mini with M5 Pro, the company says it provides 4x faster LLM prompt processing performance in LM Studio and compared to the M2 Pro, it’s 8.5x faster.

Even if you’re not interested in AI, the Mac mini with M6 and the Mac mini with M5 Pro will make for great desk companions and a great alternative to one of the best MacBooks for those who prefer a more stationary setup.

Regardless of which new Mac mini you pick up, you’re getting Wi-Fi 7 and Bluetooth 6 support as well as an upgraded 2.5 Gigabit Ethernet port around back. Both mini PCs will also ship with macOS 27 Golden Gate, which includes the next version of Apple Intelligence and a first for the Mac mini, Siri AI.

Pre-orders start today, with units scheduled to start arriving on September 22. Here’s everything you need to know about Apple’s new Mac mini.

Mac mini: Price & Availability

The Mac mini with M6 starts at $899 (or $799 for education), while the Mac mini with M5 Pro starts at $1,699 (or $1,599 for education). Apple will also offer configure-to-order options on its site where you can bump up the unified memory in the Mac mini with M6 from 16GB to 32GB and the Mac mini with M5 Pro all the way up to 64GB.

Pre-orders begin today on Apple’s website and in the Apple Store app in 30 countries and regions, including the U.S. Units begin shipping on September 22.

Mac mini: Design

The new Mac mini with M6 and Mac mini with M5 Pro are both compact mini PCs that are small enough to find a place on any desk. Apple hasn’t announced any major exterior changes, so it’s likely that both will feature the redesigned form factor first introduced with the Mac mini M4.

This means there will be two USB-C ports and a headphone jack on the front of both Mac mini models. Around back is where things differ. While the Mac mini with M6 will retain the three rear Thunderbolt 4 ports of its predecessor, the Mac mini with M5 Pro will have three rear Thunderbolt 5 ports instead. Likewise, both models will have HDMI around back along with upgraded 2.5Gb Ethernet out of the box, though you can upgrade either to a 10 Gigabit Ethernet port, albeit in a custom configuration.

On the Mac mini with M5 Pro, its Thunderbolt 5 ports deliver up to 120Gb/s which is useful for connecting fast external storage and monitors like the Apple Studio Display and Apple Studio Display XDR. However, there’s another benefit to having these faster ports. With Thunderbolt 5, you’ll also be able to cluster multiple Mac mini systems together which allows you to run large AI models locally instead of over the cloud.

Wireless connectivity is getting a big upgrade on both new Mac mini models too. Instead of Wi-Fi 6E and Bluetooth 5.3 on the Mac mini M4, the Mac mini with M6 and the Mac mini with M5 Pro now support Wi-Fi 7 and Bluetooth 6. When paired with one of the best Wi-Fi 7 routers, you’ll likely see faster downloads and uploads while the addition of Bluetooth 6 brings precise device tracking, lower latency and better power efficiency for your wireless peripherals like Apple’s Magic Mouse, Magic Keyboard and Magic Trackpad as well as your AirPods.

Although Apple didn’t reveal how many displays these new Mac mini models can drive, you’ll likely be able to use three of the best monitors simultaneously, though their resolutions and connection configurations will likely differ depending on whether you’re running the all-new M6 chip or the more powerful M5 Pro.

On the video front, we’re also getting a brand new feature delivered through USB-C called genlock support. Available on both the Mac mini with M6 and Mac mini with M5 Pro, it enables precise synchronization between a display and camera, including the iPhone 17 Pro.

Finally, storage speeds on both Mac minis are getting an upgrade. Apple switched to a next-generation SSD architecture based on PCIe Gen 6 for its upgraded Mac Studio and it looks like the new Mac mini will be following suit. The company claims up to 2x the performance compared to prior generations, which should make loading projects, transferring files and working with large language models even faster.

Mac mini: Specs & performance

Just like with the last generation, there will be two distinct versions of the Mac mini: one designed for consumers and one for pros. While both previously used different versions of the same chip, this time around, Apple is mixing things up.

The base model will be powered by Cupertino’s all-new M6 chip which features a 12-core CPU alongside a 12-core GPU. Single-threaded performance is the big upgrade here and Apple claims that the M6 offers the “world’s fastest,” so expect everything to feel extra snappy and responsive. Likewise, we’re also getting Neural Accelerators inside each core of its 12-core GPU which is a first for Mac mini. As a result, Apple claims this new chip will deliver up to 4x faster AI performance and 2x faster graphics. Multitasking is getting a boost too thanks to higher memory bandwidth up to 170GB/s.

The Mac mini with M5 Pro takes things even further performance-wise for professionals who aren’t ready to step up to a Mac Studio. With up to an 18-core CPU and up to a 20-core GPU, this upgraded model redefines what’s possible on a small desktop to handle tasks like app development, video rendering, scientific simulations and more. You’re also getting third-generation ray tracing for even more complex 3D design, VFX and game development.

With up to 64GB of unified memory with 307GB/s of memory bandwidth, the Mac mini with M5 Pro looks to be a tiny AI powerhouse capable of running even larger local AI models. However, even with this extra power under the hood, it remains a quiet and efficient always-on desktop that’s ideal for creative workflows or running AI agents.

Mac mini: Outlook

The new Mac mini is following in the footsteps of its predecessor, both in terms of design and use case. While the Mac mini with M6 will be a great desk companion with upgraded performance and improved local AI capabilities, the Mac mini with M5 Pro is more of a workstation for professionals who want even more performance in the same small form factor.

Both Mac mini models are getting a bit of a price hike compared to the previous generation but with higher memory and storage prices, this is to be expected. We’ll see what the Mac mini with M6 and the Mac mini with M5 Pro are truly capable of when we get them into our testing lab in September, so stay tuned for more.

If AI is supposed to replace our jobs and make everyone's lives better what happens when it's put into robots? The CEO of German robotics company Neura Robotics believes that physical AI could replace human labor in less than five years.

Speaking with Tom Mackenzie on Bloomberg, David Reger, the founder of Neura Robotics, called the pace of physical AI development and robotics as "tremendous."

The discussion is mainly about how robotics can be a growth sector for European countries. The focus of the interview is on productive robotics and the opportunity they represent especially in the future.

"I think physical AI and robotics will take every physical tasks," Reger said. "I'm not talking in the next 10 years, I'm talking in the next two to five years."

Reger went on to say that he can train a robot in a task faster than he can train a human.

On a consumer level, robots are mostly relegated to lawnmowers and vacuum cleaners. But on an industrial level we can already see where robotics is growing, especially on assembly lines like those that build cars.

However, a push for vast number of humanoid robots is coming. The Japanese company Noetra (not to be confused with the German Nuera), has a grand plan to deploy more than 10 million humanoid robots across various industries by 2040, and its backed by the Japanese government.

These would be inserted into key areas like disaster response, manufacturing, eldercare and the decommissioning of the Fukushima Daiichi Nuclear Power Plant. Earlier this year, doctors and engineers in San Diego performed the first two surgeries assisted by humanoid robots.

Consumer robots are coming

On a base level, robotics are out of reach for most people. Per RoboZaps, the average cost of a humanoid robot is well into the 10s of thousands of dollars. Most cost as much as a new car and are meant for research, developers or industrial deployments.

That said, Nor Robotics out of San Francisco just announced its first sub-$2k Nori A3 robot built in the United States, as reported by Forbes. The Nori A3 is being built and sold with the explicit goal of making advanced home robotics more accessible.

"We wanted to build the best robot possible for under 2K and in the US. With this version Nori we think we've finally done it," CEO Antonio Li said on X.

The Nori A3 can lift up to 122 pounds vertically, has six to eight hours of battery life, and 19 degrees of freedom (or movement). Allegedly, it can pour a drink, load the dishwasher or fold clothes.

For $1,600 would you take a chance on putting a robot in your home?

More from Tom's Guide

• Meet Moya, the AI that winks, smiles and walks with 92% human-like gait — is this the uncanny valley?
• I tried a robot lawnmower for the first time — and there's one thing I wish I'd known before using it in my yard
• 'The future will be a mix of bionic and humanoid robots': Dreame’s product expert reveals the brand's next step for robot vacuums

Apple's September event is here, and it's likely to bring us a new pair of AirPods, either the recently rumored AirPods 5 or the entirely new, cutting-edge AirPods Ultra.

Reporting from folks like Mark Gurman at Bloomberg claim that the so-called AirPods Ultra will leverage a new H3 chip, built-in cameras and Apple Intelligence to detect objects around you. They'll be a step in between the AirPods Pro 3 with heart-rate detection and more intrusive wearables like Meta's "pervert glasses."

So, what exactly do we know about these futuristic earbuds? Here's a full breakdown.

AirPods Ultra: Potential release date

The earliest reports of a camera-equipped pair of AirPods came from noted Apple analyst Ming-Chi Kuo. In a report dated July 1 2024, Kuo said that "the cameras can detect changes in the environment" and could "potentially enable in-air gesture control to enhance human-device interaction."

The legend of the camera-equipped AirPods Ultra has only grown from there.

These days, the person who seems to know the most about the AirPods Ultra is Bloomberg's Mark Gurman. Back in January of 2025, Gurman said in a newsletter that Apple was “exploring other types of wearable products” that included a pair of AirPods that could come with miniature infrared cameras. Back in June, he said that we're likely looking at a 2027 release date. This was then repeated in a post on X.

It's possible Apple might unveil them at this September's iPhone 18 event, but I'm not entirely convinced. Apple is rarely in a rush to release a product, especially one as potentially complex as the AirPods Ultra are rumored to be, so I'm not exactly putting money on this.

The most likely scenario based on the reports I've read peg the AirPods 5 as the most likely candidate for the 2026 event and the AirPods Ultra for a Spring or Fall 2027 launch.

AirPods Ultra: Potential features

So what are the AirPods Ultra going to do different from the current-gen AirPods Pro 3? Well, it turns out, quite a lot.

As far as we know right now, there are two different models in the works: There's a B790 model and a B798 model. The former is closer to launch than the latter, and it's unlikely that both will debut at the same time.

On the outside of the earbuds will be IR cameras that can scan objects around you using Apple's Visual Intelligence. (If you have long hair that impedes the view of one of these cameras, the earbuds will ask you to move it to the side.)

A video purportedly found in the latest macOS Tahoe 26.7 software shows what this could look like in action.

The AirPods Ultra's IR cameras are, reportedly, lower resolution than Meta's AR glasses and will feature a light-up LED that will turn on when it's scanning, which should mean that they'll be less of a privacy concern than other wearables we've seen lately.

There are a number of potential uses for IR-enabled earbuds, including proximity detection to other people that can improve ANC functionality.

To power this functionality would be a new H3 chip. Little is known about it other than the fact that it is the key driver of the deeper Apple Intelligence integration. Better processors usually enable better noise cancelling, however, so expect to see that as a feature Apple touts sometime in the future, too.

AirPods Ultra: Potential price

All this cutting-edge tech isn't likely to come cheap. A fair assumption right out of the gate is that the AirPods Ultra will cost more than the current-generation AirPods Pro 3 that currently sit at $249 / £249 / AU$429.

I would not be surprised if Apple charged $349, which would be $100 more than the AirPods Pro 3. There is some wiggle room if Apple wants to increase the sticker price of its new earbuds: Two of the AirPods' biggest rivals, the Sony WF-1000XM6 and Bose QuietComfort Ultra Earbuds 2, cost $299.

For now, all we can do is prognosticate what Apple will do with pricing as the AirPods Ultra are truly a one-of-a-kind device that doesn't really match up with anything on the market at the moment.

AirPods Ultra outlook: Why I'm skeptical

The AirPods Ultra do sound like an extremely inventive innovation in the world of wireless earbuds. The ability for a pair of earbuds to detect objects around a user has major implications for the future of wearable technology.

However, I have two major reservations about the AirPods Ultra.

The first is that it's unlike Apple to jump first when it comes to launching a new type of product. Apple wasn't the first AR glasses manufacturer, for instance, and it certainly wasn't the first to make an MP3 player or laptop. What Apple usually does is that it lets competitors lead and then presents a more elegant solution.

Even if the AirPods Ultra are production-ready, that doesn't mean Apple will put them out without waiting for something else on the market first.

I'm also skeptical about the cameras themselves. Apple paints itself as extremely privacy focused. Putting cameras on a pair of earbuds flies in the face of that stance. Unless there are some serious limitations of the technology (like the only data it collects are IR images), it'll be hard for Apple to keep its privacy-first reputation.

But if Apple finds the balance between features and privacy, locks in at a reasonable price and gets developers on board, the AirPods Ultra could truly be the next big thing in wireless earbuds.

More from Tom's Guide

• AirPods Ultra with cameras are 'nearly ready' — here's 4 features you need to know
• AirPods Ultra 'almost certainly' hinted at in iOS 27 code — and they could be a huge audio (and camera) upgrade
• Apple iPhone 18 event — all the new products expected to launch in September

The Wall Street Journal says Apple is talking to publishers about multi-year deals. The idea is to let Siri AI use their reporting to answer questions about what’s happening now. Perhaps it will look something like Gemini AI is already doing with Google, but Apple’s budget for this but be even bigger. According to the WSJ report, possibly hundreds of millions, depending on how much of each publisher’s work Siri ends up using.

Actually paying publishers is a shift from the usual scrape-now, argue-later routine that’s defined most of the AI industry so far. It means Siri might finally get real reporting and fewer confidently wrong answers.

Apple is betting on a new model of the news business

Apple seems to be betting on a version of the news business that doesn’t really match how most people get their information anymore. For instance, Siri can’t just parrot whatever’s trending on X, TikTok or YouTube. Apple needs sources with editors, corrections and someone who actually takes the blame when things go wrong.

The company has already learned how quickly AI-generated news can create trouble. Apple suspended notification summaries for news and entertainment apps in early 2025 after Apple Intelligence produced inaccurate summaries of publishers’ alerts. In one widely reported example, it falsely made it appear that the BBC had reported that murder suspect Luigi Mangione had shot himself.

Licensing real reporting means Apple can control what Siri says and how it uses that info. Publishers get paid, which matters now that AI answers are making it less likely anyone clicks through to the original version story.

For questions such as “What did the Federal Reserve announce?" If you ask Siri what the Fed did today or what’s up with the election, it’s safer to pull from vetted sources than to let Siri riff on whatever’s blowing up on online.er since audiences are finding news somewhere else.

People are consuming news through multiple outlets

How people actually find news has changed a lot faster than the newsrooms making it. The Reuters Institute’s 2026 Digital News Report found that social media and video networks have overtaken publishers’ websites and apps as a source of news across many of the markets it studied. Globally, 30% of respondents now describe social and video platforms as their main source, up from 22% five years ago.

The difference is even sharper among younger people. More than half of 18- to 24-year-olds said social media, video platforms and AI chatbots were now their primary route to news. For people between 18 and 34, the use of news websites and apps has fallen so quickly that those properties no longer represent the leading source for any age group globally.

Where those users place their attention has changed as well. A separate Reuters Institute study of young news audiences found that 51% of social media news users between 18 and 24 pay the most attention to individual creators and personalities. Only 39% said the same about traditional news organizations or journalists.

The pattern is visible in the U.S., too. Pew Research Center found that 38% of Americans ages 18 to 29 regularly get news from influencers, compared with 8% of people 65 and older. More than three-quarters of those younger users said they usually encountered that coverage rather than deliberately searching for it.

Even the safest Siri could still feel out of step

None of this means Apple should train Siri to repeat a podcaster’s speculation or treat an unverified video as established fact. Independent reporting and creator commentary vary enormously in quality, and social platforms can spread false information much faster than a newsroom can correct it. But the danger is at the other extreme. If Siri leans too hard on several of big publishers, the answers might be technically correct but still miss the point. It could repeat what’s in the official record and skip the interviews, eyewitness clips, niche newsletters and independent reporting that actually make for good reporting.

Without a doubt, this will make Sirie end up sounding disconnected with a summary that can't confirm the difference between firsthand reporting, analysis and random speculation. All of which, real human reporters can typically discern.

So while publishers might be able to give Siri a solid base, but they can’t cover the whole news ecosystem on their own.

Amanda's take

As a journalist, I want publishers to be paid when AI companies use their reporting. Apple licensing credible news is far better than Siri scraping articles, stripping away the context and answering without sending anything back to the people who did the work.

But I don’t think established publishers alone reflect how people understand the news anymore. I read traditional reporting, but I also listen to podcasts, follow independent journalists and watch firsthand videos that may capture a story before a major newsroom catches up. Younger iPhone users are even more likely to piece together what’s happening that way.

That doesn’t mean Siri should treat every viral post as fact. It means Apple needs a smarter system than a short list of approved publishers. Siri should be able to begin with what has been verified, clearly identify what remains unconfirmed and acknowledge credible reporting or firsthand accounts emerging elsewhere.

Apple is right to make reliability the foundation. But if it plays things too safely, Siri could end up giving accurate answers that still feel a day behind the conversation.

The Siri I want would not force me to choose between established journalism and the people documenting events in real time. It would help me understand which claims are confirmed, where they came from and why people are talking about them. Paying publishers is a strong start, but it cannot be Apple’s entire answer to how news works now.

More from Tom's Guide

• I put ChatGPT on my Ray-Ban Meta glasses — and it completely changed how I used them
• Claude is now watermarking every response — here's what that means if you use AI for writing
• 5 custom Gemini Gems that I swear save me hours each week

Anthropic recently announced that all new Claude models will embed an invisible signal into their text output. The goal is to help identify AI-generated content and meet new transparency requirements under the European Union's AI Act.

But if you're picturing an invisible digital stamp that permanently proves the words were written by AI, that's not exactly how it works. Instead, Claude subtly changes the statistical patterns in its writing. The text still reads naturally to humans, but software designed to look for those patterns can potentially recognize that it came from Claude.

The announcement has reignited debate over whether AI watermarking is finally the answer to identifying AI-generated writing or if it even works.

What is an AI watermark?

Unlike a copyright watermark on an image or hidden metadata inside a file, a statistical watermark is built directly into the words themselves. The AI slightly favors certain word choices and sentence patterns while generating text.

Individually, those choices are meaningless to us as humans, but across hundreds of words, they create a statistical signature that specialized detection software can identify. It's kind of like a fingerprint that is only recognized by software trained to understand the pattern.

The watermark itself is invisible to readers and doesn't appear anywhere in the document. Copying text into Word, Google Docs or an email won't reveal anything unusual because the signal is designed to be detected only by software that understands the statistical patterns Claude leaves behind.

Does watermarking affect the quality of Claude's writing?

Slightly, but let's be real, AI writing isn't great to begin with. Alexander Cui, head of research at GPTZero, tells Tom's Guide that the statistical watermarking does introduce a small tradeoff because the model has to make slightly different word choices than it otherwise would.

"The output quality does subtly decrease," Cui said. "Directionally, it does reduce the output quality."

He added that modern watermarking techniques have become sophisticated enough that most users are unlikely to notice the difference in everyday use.

And, while the watermark can be removed, it makes things much more complicated. Editing a few words won't remove a statistical watermark, in fact, many watermarking systems are designed to survive minor edits or paraphrasing.

But the signal becomes much more fragile if Claude's response is rewritten by another AI model. That is, if you asked Claude to write your science paper and then asked ChatGPT to revise the draft. In that case, the watermarking becomes important because Claude's output often isn't the final product. A response might become the input for another chatbot, an AI coding agent or a document editing assistant before anyone reads it.

Cui says the watermark generally doesn't survive that kind of workflow.

"The watermark applies very subtle changes," he explained. "They usually would not affect downstream output. What you need is the actual text written by the model itself."

Does this mean AI detectors finally have proof?

Anthropic's watermark is one additional signal, but it's not definitive proof that a document was written by AI.

Alon Yamin, CEO of AI detection company Copyleaks, says statistical watermarks can help identify AI-generated content, but shouldn't be treated as a silver bullet.

"While this seems like a big step forward for AI detection on the surface, it's far from a silver bullet," Yamin said.

He argues that watermarking should be considered alongside other evidence rather than replacing broader AI detection methods.

These days, modern documents often mix human writing with AI assistance. From tools like Grammarly to help edit or Apple Intelligence Writing Tools, these common AI features have become integrated into daily use by many. For example, someone might draft an article themselves and ask Claude to improve grammar, reorganize paragraphs or rewrite a headline.

Does ChatGPT do the same thing?

OpenAI has spent years researching watermarking and other provenance technologies, but it has not broadly deployed statistical text watermarking across ChatGPT responses.

Google has also explored watermarking techniques for AI-generated content, while several companies have experimented with metadata-based approaches instead of statistical signals embedded in text.

As a result, Claude is currently taking one of the most visible steps toward model-level watermarking for everyday AI writing.

You'll likely never notice a watermark while using Claude, and it won't suddenly make AI-generated writing easy to identify. But where it could matter is in education, publishing, enterprise software and workplaces that need to disclose when AI was used to create content.

Bottom line

The big picture here is that watermarking is a tool companies are experimenting with as governments push for greater transparency around AI-generated content. It can be useful under the right conditions, disappear after enough editing or rewriting and works best when combined with other evidence rather than on its own.

But as of August 2, 2026, every word Claude writes now contains a hidden statistical watermark that most people will never see.

More from Tom's Guide

• I put ChatGPT on my Ray-Ban Meta glasses — and it completely changed how I used them
• ChatGPT can do more for you than ever — and I use these 5 prompts to take everyday tasks off my plate
• 5 custom Gemini Gems that I swear save me hours each week

I've spent years tweaking my iPhone, but after seeing yet another viral "privacy screen" hack making the rounds online, I wondered what I was still missing.

So instead of copying another internet trick, I asked ChatGPT to audit my iPhone's privacy settings. Here's the prompt I used:

"I'm an iPhone user who often works in coffee shops, takes weekend getaways and occasionally hands my phone to family members. What privacy settings should I enable to better protect my personal information? Include settings that are easy to overlook and explain why each one matters."

Some of ChatGPT's recommendations were obvious. Others were buried deep inside iOS — and a few were settings I'd simply never taken the time to enable.

After trying every suggestion myself, these are the ones I'd recommend to almost anyone.

1. Hide sensitive notification previews

This was ChatGPT's first recommendation, and it's probably the easiest privacy win on the list.

Without changing anything, your iPhone can display text messages, banking alerts, two-factor authentication codes and other notifications right on your Lock Screen. Anyone nearby can glance at them without ever unlocking your phone.

Where to find it: Settings > Notifications > Show Previews

The tweak: Change the setting to When Unlocked.

Changing previews to When Unlocked means Face ID has to authenticate you before those details appear.

My take: I can't believe I left this turned off for so long. If you ever leave your phone face-up on a desk or café table, this setting alone is worth changing.

2. Lock individual apps with Face ID

One of Apple's best recent privacy features lets you require Face ID before opening supported apps. And I don't just mean your banking information or credit card apps.

Whether it's Notes, Photos or anything else containing sensitive information, this adds another layer of protection if someone else picks up your phone.

Where to find it: Long-press an app on your Home Screen > Require Face ID

My take: I started locking Notes and Photos immediately. They're probably the two apps I'd least want someone casually browsing.

3. Review which apps know your location

While checking my own phone, I found several shopping and restaurant apps that still had broader location access than they actually needed. Switching them to While Using the App took only a few minutes.

Where to find it: Settings > Privacy & Security > Location Services

ChatGPT suggested reviewing location permissions instead of leaving everything set to Always.

My take: This is the kind of setting that's easy to forget once you've tapped "Allow" the first time. Now that I've taken a look at what apps had my location, I recommend checking this periodically.

4. Turn on Safety Check

Safety Check is one of Apple's least-known privacy tools. Originally designed for people who need to quickly revoke shared access during unsafe situations, it also gives anyone an easy way to review who can see their location, shared photos and other permissions.

Where to find it: Settings > Privacy & Security > Safety Check

My take: Even if you never need the emergency features, it's a great way to see what information you're sharing without realizing it.

5. Audit camera and microphone permissions

It turns out I had granted camera and microphone access to apps I hadn't opened in months. Removing permissions only takes a few seconds, and it helps limit which apps can access some of your most sensitive hardware.

Where to find it: Settings > Privacy & Security

My take: This one surprised me the most. I hadn't realized how many apps were still hanging onto these permissions.

6. Hide apps you don't want anyone else opening

If you have apps containing financial information, medical records or personal documents, you can hide them completely. The app disappears from your Home Screen and is moved into a locked Hidden folder inside the App Library.

Where to find it: Long-press an app > Require Face ID > Hide and Require Face ID

My take: This is perfect if you occasionally hand your phone to your kids or someone asks to borrow it for a minute.

7. Turn on App Privacy Report

App Privacy Report shows a rolling log of how often apps access your location, camera, microphone and network connections. You might be surprised how active some apps are, even when you rarely use them.

Where to find it: Settings > Privacy & Security > App Privacy Report

My take: Privacy isn't something you configure once and forget. Checking this report every few weeks makes it much easier to spot apps that are asking for more than they really need.

What ChatGPT missed

Overall, ChatGPT's recommendations were solid, but it overlooked two of Apple's most useful security features.

The first is: Turn on Stolen Device Protection

Where to find it: Settings > Face ID & Passcode > Stolen Device Protection

If someone learns your passcode and steals your phone, this feature adds extra safeguards before critical account changes can be made. When you're away from familiar locations like home or work, iOS requires Face ID for sensitive actions and places a security delay on certain account changes. It's one of Apple's best anti-theft features, and I wouldn't leave it turned off.

The second feature is: Disable Control Center on the Lock Screen

Where to find it: Settings > Face ID & Passcode > Allow Access When Locked

Consider turning off Control Center access from the Lock Screen. Doing this makes it harder for someone who steals your phone to immediately change connectivity settings before you've had a chance to locate your device.

If you rarely use Control Center without unlocking your phone first, it's a small convenience to give up for some added peace of mind.

One Apple Intelligence setting that's worth checking

If you use Apple Intelligence, there's one extra privacy consideration that didn't come up in ChatGPT's recommendations.

Apple Intelligence can summarize long emails, messages and notifications to help you catch up more quickly. That's useful, but it also makes hiding Lock Screen previews even more important.

If notification previews are visible, those AI-generated summaries may reveal more information at a glance than the original notification would have.

Apple has taken a privacy-first approach to Apple Intelligence by processing many requests directly on your device and using its Private Cloud Compute system for more demanding tasks. But no amount of on-device processing can protect information that's already visible on your Lock Screen.

That's why changing notification previews to When Unlocked was the very first setting I kept enabled after testing ChatGPT's suggestions.

Final thoughts

The viral privacy screen hack that inspired this experiment didn't make the cut. I don't think I need one after enabling these settings. Sure, if you're worried about shoulder surfers on planes or in coffee shops, a physical privacy screen protector is still your best bet.

But what surprised me most was how many of ChatGPT's suggestions I had missed despite using an iPhone every day. It gave me a useful checklist, but the real value came from taking 15 minutes to review what my phone was actually sharing.

If you haven't looked through your iPhone's privacy settings in a while, it's worth doing. You may discover, like I did, that a few small changes can make your device feel a lot more secure.

More from Tom's Guide

• I found the 7 easiest Claude connectors for people over 50 — start with these
• I replaced my notebook with an AI pen for a month — here's what happened
• Looking for a job after 50? Here are 7 ChatGPT prompts to level the playing field

Apple is expected to launch iPadOS 27 this Fall, but you can try the upcoming operating system right now via the public beta. While the OS is still a work in progress, it has most of the essential new features coming to the best iPads.

Like iOS 27 and macOS 27, iPadOS 27 is focused on overall system refinements and deeper AI integration. That last bit is especially noteworthy, as there is now a dedicated Siri AI app that functions like a standard chatbot. Beyond answering questions, Siri AI can also take actions on your behalf since it’s integrated into the operating system and works with Apple’s applications.

Though iPadOS 27 isn’t a monumental update, features like the overhauled Siri AI, the Liquid Glass slider, and other quality-of-life improvements make it worthwhile. Here are my impressions of the upcoming iPad operating system and its new features.

iPadOS 27 hands-on preview: Compatible iPads

Here are all the iPads that support iPadOS 27.

• iPad Pro (M4 and later)
• iPad Pro 12.9‑inch (4th generation and later)
• iPad Pro 11‑inch (2nd generation and later)
• iPad Air 13-inch (M2 and later)
• iPad Air 11-inch (M2, M3, and M4)
• iPad Air 11-inch (4th generation and later)
• iPad (A16)
• iPad (9th generation and later)
• iPad mini (A17 Pro)
• iPad mini (6th generation)

iPadOS 27 hands-on preview: New features

Siri AI

Apple Intelligence is finally ready for prime time on the company’s devices. Thanks to the new Siri AI app, you can now interact with Apple’s virtual assistant the same way you would with a chatbot like ChatGPT or Google Gemini. Apple Intelligence runs on both servers and your device to offer better image understanding and more.

There is now a dedicated Siri AI app, which looks and functions like a regular chatbot app. In this window, you can ask Siri questions or have it perform tasks for you. For example, I used Siri to send a message to one of my iMessage group chats, and it did so after my approval. Sure, I could have just done this directly in the group chat, but it’s a good demonstration of what the new Siri can do.

Siri AI also understands what it sees on your screen. When I had the Tom’s Guide homepage open on Google Chrome, it correctly detailed what was on the screen, including all the articles. It couldn’t open the articles or read them out loud (despite saying it could), but given that I’m testing a beta version, I’ll give it an A for effort, even if it failed to complete the task.

Siri is most useful when you’re using Apple’s own apps and programs. It can create events on your calendar, set reminders, play music, and more. For example, when you’re looking at a schedule, you’ll see an “add to calendar” button pop up. When you right-click certain folders, files, or apps, you’ll see an “Ask Siri” option at the bottom. Selecting this will open a window where you can start texting or talking with Siri.

Though Siri can’t yet interact with most third-party applications (such as those from Google), it’s a huge improvement over the old Siri.

Liquid Glass improvements

Liquid Glass hasn’t gone anywhere, as Apple is leaning more into the UI with its latest operating system updates. However, Apple is now giving you greater control over how it looks thanks to a new slider that adjusts the opacity. You can’t eliminate the Liquid Glass look, but you can severely minimize it.

There are other improvements, such as how complex background visuals now appear more diffuse so that text stands out more. You’ll also notice that the edges of icons are more pronounced to help further separate them from the background. The uniform corner radius of all app windows also helps them stand out more.

Even if Liquid Glass hasn’t gone anywhere, even detractors should find the updates tolerable since they do greatly improve overall readability and visibility. The Liquid Glass slider is a nice touch.

I've never had a problem with Liquid Glass because I think it’s an interesting visual effect. That said, I’m glad that folks can now fine-tune it to their liking. The slider is especially neat because it’s easy to use and makes it clear how much of a difference there is when you go from one extreme to the other.

Child safety features

If you’re a parent, you’ll be happy to know that Apple is expanding its parental control features with iPadOS 27. This way, you can let the kiddo use Apple’s tablet without worrying too much about the kind of content they’re watching and for how long they’re using the slate.

Some of these updates include a redesigned and more intuitive Screen Time Dashboard that offers one-tap controls and expanded safeguards. There is also Time Allowances for app categories, Ask to Browsw website approval, Communication Safety that blurs gore or violence, and one-tap access management. You can tailor these features to suit your child.

I don’t have kids but I did find the child safety features simple enough to use and customize. Given how much kids love iPads, it’s smart for Apple to continue making adjustments to these safety features.

iPadOS 27 hands-on preview: Other features

Spatial Reframing: As the name suggests, you can now reframe a photo after it’s been taken thanks to this feature. You can also expand your shots with the Extend Tool and remove objects with the Clean Up tool.

Upgraded Image Playground: Image Playground can now generate higher quality images than before. You can also modify and transform existing images by describing what you want.

Siri in Notes: Siri AI can assist with your handwritten or typed notes in the Notes app by (for example) organizing bullet points into a formatted agenda or turning handwritten lectures into a detailed study guide.

Safari updates: In Safari, tabs can now be automatically grouped by topic to make related pages easier to find. With Safari Notify Me, you can receive an alert if there’s a price change or a restock for an item you’ve been eyeing.

Describe Shortcuts: Instead of manually building complex workflows, you can now describe what you need, and Shortcuts connects actions across multiple apps to make it happen. This is useful if you want to set specific routines, such as launching a specific app when you connect a peripheral.

iPadOS 27 hands-on preview: Outlook

iPadOS 27 won’t set the world on fire with revolutionary new features, but the upgraded Siri AI and deeper Apple Intelligence integrations help make iPads that support the update more useful. This is especially true if you frequently use Apple’s own apps.

Keeping in mind that iPadOS 27 is still a work in progress, I think it’s a smart update that finally sees Apple’s AI ambitions bear fruit. Things could change between now and when the expected Fall launch, but right now, Apple has delivered on its promises for a more helpful and intuitive iPad experience.

More from Tom's Guide

• Apple hid a perfect 'dumb phone' feature inside your iPhone that's great for kids — here's how to turn it on
• iOS 27 beta now lets you customize Siri's pace and expressivity — here's how to do it
• iOS 27's new AI feature will automatically fix your weak passwords — how to set it up

You don’t have to wait to try iPadOS 27, as you can try it out now via the public beta. Even if this version of the operating system isn’t groundbreaking, its emphasis on AI should make for a more personal and customized experience.

As with iOS 27 and macOS Golden Gate, the main attraction is the reinvented Siri, which finally lives up to its potential. Siri AI, as it’s now called, is context-aware and understands what’s on your screen. Because it’s built into the OS, it can directly control and manage apps. There are also new child safety features, improved readability for Liquid Glass, new photo-editing tools, and more.

While public betas are generally safer than developer betas, it’s best to proceed with caution, as there may be bugs that could harm your iPad. If you’re not bothered by that (I’ve never had issues in all my years of running these Apple betas), then here’s how to download the iPadOS 27 public beta.

How to download the iPadOS 27 public beta

Before you install the beta, make sure to back up your iPad, just in case anything goes wrong and you need to return to iPadOS 26.

You’ll then need to enroll in the Apple Beta Software Program. You may already be enrolled from prior years, but if not, visit the sign-up page to get the ball rolling. You’ll need to log in with your name and address, then set your entity type as an individual. When the payment screen appears, back out, and you’ll be on the free developer tier.

Here is the full list of iPads that are compatible with iPadOS 27:

• iPad Pro (M4 and later)
• iPad Pro 12.9‑inch (4th generation and later)
• iPad Pro 11‑inch (2nd generation and later)
• iPad Air 13-inch (M2 and later)
• iPad Air 11-inch (M2, M3, and M4)
• iPad Air 11-inch (4th generation and later)
• iPad (A16)
• iPad (9th generation and later)
• iPad mini (A17 Pro)
• iPad mini (6th generation)

And that's how you download the iPadOS 27 public beta! You should be safe using the beta, but if you don’t want to risk it, you can always wait for the full release this Fall.

More from Tom's Guide

• I just downloaded macOS 27 public beta — skip Siri AI and try these 5 hidden features instead
• Haptic vs. Mechanical trackpads: I put them both through a gauntlet, and one almost made me throw my laptop into the sun
• Exclusive: I asked Lenovo’s GM to predict the future of monitors, and it's the end of our chaotic, cable-cluttered desks and monitors that fry your retinas

The macOS Golden Gate public beta is now live, giving you a chance to check out all the new features coming to the best MacBooks.

Expected to launch this Fall, macOS Golden Gate primarily focuses on updating Apple Intelligence and Siri, with the latter now officially called Siri AI. You don’t get many new features, but improvements to Liquid Glass’s readability, easier customization, and smoother performance make for an overall better experience.

While Apple’s public betas are generally more stable than the developer builds, it might still be buggy and could cause issues for your Mac. I’ve never experienced serious issues using any of these betas over the years, but it’s worth being cautious. With that said, here’s how to download the macOS Golden Gate public beta.

How to download the macOS Golden Gate developer beta

Downloading the macOS Golden Gate public beta is pretty simple. First, enroll your Apple ID in Apple's Beta Program and make sure your Mac is compatible. Here's the full list of Macs that can install and run macOS Golden Gate:

• MacBook Neo (2026)
• MacBook Air with Apple silicon (2020 and later)
• MacBook Pro with Apple silicon (2020 and later)
• iMac with Apple silicon (2021 and later)
• Mac mini with Apple silicon (2020 and later)
• Mac Studio with Apple silicon (2022 and later)
• Mac Pro with Apple silicon (2023 and later)

That’s all that’s required to download the macOS Golden Gate public beta. Your Mac will restart once the download finishes. After that, you can start trying out all the new features. Depending on your connection speed, the download might take a few minutes, so be patient.

More from Tom's Guide

• I just downloaded macOS 27 public beta — skip Siri AI and try these 5 hidden features instead
• Haptic vs. Mechanical trackpads: I put them both through a gauntlet, and one almost made me throw my laptop into the sun
• Exclusive: I asked Lenovo’s GM to predict the future of monitors, and it's the end of our chaotic, cable-cluttered desks and monitors that fry your retinas

We've all been there. You're sitting in a dimly lit restaurant, squinting at a menu with tiny print, or standing in the grocery store trying to figure out whether the yogurt in your hand expires in three days or three weeks.

I've worn glasses since I was three years old, so I know the struggle. Without them, I've been known to snap a photo of a label just so I can zoom in and read it.

But there's an easier way than squinting and hoping for the best.

One of the most useful accessibility features on the iPhone is tucked away inside the Magnifier app. Using Apple's on-device AI and computer vision, it can detect text, identify objects, describe your surroundings and even read printed text aloud in real time. Once you start using it, you'll probably wonder how you ever managed without it.

How to use your iPhone's AI magnifier

1. Open Magnifier. Open the Magnifier app on your iPhone. If you can't find it, swipe down from the Home Screen and search for Magnifier.

2. Turn on Scene Description. Tap the gear/settings button and make sure Scene Description is added to your controls. If prompted, download the required Apple Intelligence models.

3. Point your camera at tiny writing such as a menu. Aim your camera at the menu, medicine bottle, receipt or any printed document.

The AI analyzes what it sees and begins describing the scene. It also simply magnifies for easy reading. You can also zoom in, freeze the image or increase contrast if you want to inspect something more closely.

4. Listen as your iPhone reads the text. Tap Scene Description. Your iPhone can read the printed text aloud while also describing what's in view, making it much easier to understand menus, signs or labels in poor lighting.

Things to know

While menus or lablels are probably the most relatable example, I've found this feature handy for plenty of everyday situations:

• Reading medicine bottles
• Checking cooking instructions on food packaging
• Reading restaurant receipts
• Looking at thermostat displays
• Examining serial numbers
• Reading labels in grocery stores
• Checking instructions on appliances

Scene Description is designed as an accessibility feature, but it's useful for anyone who occasionally struggles with small print.

The feature works best in good lighting and with clearly printed text. If you're trying to read heavily stylized fonts or handwritten notes, results may vary. You'll also need a compatible iPhone that supports Apple Intelligence, and some models may require downloading the on-device AI models before using the feature.

The takeaway

This is one of those features that flies under the radar, but once you discover it, you'll want to tell all your friends. I use Scene Description almost as much as other features like Writing Tools or Genmoji.

For me, it's one of the most practical AI features on the iPhone. Whether you're trying to decipher a restaurant menu, read a prescription label or make out tiny instructions on product packaging, it can save you from squinting or hunting for your reading glasses.

Give it a try and share your thoughts in the comments.

More from Tom's Guide

• Someone created a 'Ghost Font' that humans can read but AI can't — I had to try it for myself
• The end of forgetting: Why AI’s perfect memory is a psychological trap
• I just found a free way to watch YouTube without ads — and it doesn't require YouTube Premium

After Apple’s dramatic showing at WWDC 2026 last month, I’ve been getting cozy with all the new features thanks to the iOS 27 developer beta. Unless you’ve signed up for a developer account, the vast majority of people need to wait for the public release.

Thankfully, the wait is over. The iOS 27 public beta has launched, and supported iPhones can finally experience the latest and greatest that Apple has to offer with its phone software. What’s fascinating is that my colleagues and I felt iOS 26 left users feeling like Apple was trailing behind in the AI wars — but iOS 27 is an absolute redemption arc.

That’s because it’s taking a big leap, complementing the foundational features of the platform with AI features and tools that don’t just make life easier; they fundamentally change how you interact with your iPhone on a daily basis. I’ll tell you how I feel about these new changes and additions with my iOS 27 public beta hands-on.

iOS 27 public beta hands-on preview: Supported iPhones

Before you take a stab at trying some of the best iOS 27 features, you’ll obviously need an iPhone that can support it. Apple has optimized the software structure so well that it runs smoothly even on older hardware, which is why official compatibility extends as far back as the iPhone 11.

However, there’s a small caveat regarding which models gain support for the most advanced Apple Intelligence features and Siri AI. Because these tools require significantly more processing power, only the last few generations of hardware will get access to them. Here is the full list of iPhones that support iOS 27, with the models in bold indicating they’ll support the most advanced AI features:

• iPhone 17
• iPhne 17e
• iPhone Air
• iPhone 17 Pro
• iPhone 17 Pro Max
• iPhone 16
• iPhone 16 Plus
• iPhone 16 Pro
• iPhone 16 Pro Max
• iPhone 16e
• iPhone 15
• iPhone 15 Plus
• iPhone 15 Pro
• iPhone 15 Pro Max
• iPhone 14
• iPhone 14 Plus
• iPhone 14 Pro
• iPhone 14 Pro Max
• iPhone 13
• iPhone 13 mini
• iPhone 13 Pro
• iPhone 13 Pro Max
• iPhone 12
• iPhone 12 mini
• iPhone 12 Pro
• iPhone 12 Pro Max
• iPhone 11
• iPhone 11 Pro
• iPhone 11 Pro Max
• iPhone SE (both 2020 and 2022 models)

iOS 27 public beta hands-on preview: a much smarter, more reliable assistant

I’ve been running the iOS 27 developer 3 beta for a while now, and nothing continues to stand out more than Siri AI. Apple’s assistant finally gets its most substantial upgrade since its inception, acquiring new on-screen contextual awareness that makes this assistant feel far more capable than before. What’s really notable is the proactive, dynamic approach that Siri AI takes when you're using your phone.

For example, if you’re reading an email or texting back and forth in Messages, Siri AI will scan the screen and provide a contextual suggestion — like offering a one-tap option to create a Calendar invite based on a discussed time and place.

When it comes to productivity, I love using Siri AI to handle trivial, repetitive tasks. In the past, it was a total time suck to manually copy and paste text or pull hashtags from an Instagram post, but now I can simply tell Siri AI to grab them for me. It knows exactly what I’m looking at on screen and understands the intent behind my request, saving me from tedious tapping and typing.

Additionally, a dedicated Siri app now aggregates all of your previous requests in one place. It organizes your history much like the Gemini app, letting you easily search through past queries and jump right back in to follow up on them.

But as much as Siri AI makes life easier, it still has some catching up to do. Most of its deepest functionality is currently locked down to Apple's native apps. It still lacks the robust third-party app integration that would allow it to execute complex, multi-step real-world tasks — like letting me place an order for Dunkin' Donuts via the DoorDash app.

iOS 27 public beta hands-on preview: Speed Boost

On the surface, Liquid Glass still makes iOS 27 look visually appealing with its sense of depth, transparency, and layered layout. While I’m happy to see that Apple is letting users tune the intensity of the transparency effects, I’m far more impressed by how much more responsive iOS 27 feels.

You can’t fully appreciate the fluidity of iOS 27 until you see it side-by-side with iOS 26. Running two iPhone 17 models side-by-side — one with the iOS 27 beta and the other with iOS 26 — immediately reveals how much smoother the system animations are and how much faster apps launch. As you can see from the data comparison chart above, the speed improvements when launching apps from the home screen are undeniable.

iOS 27 public beta hands-on preview: Reframe and Photo Clean Up

I’m big on generative AI and how its features can enhance my photos. Fortunately, iOS 27 brings some much-needed Apple Intelligence upgrades to the table, and I for one am thoroughly impressed by what they can do.

The standout feature you need to know about is Reframe. I’ve used it countless times already to shift the perspective of photos I've captured — and they don’t even need to have been shot on an iPhone. For instance, I used it to adjust the angle of a BMW 7 Series photo to show more of its side profile than the original shot did. Honestly, it’s unlike any other AI photo tool I’ve come across on a smartphone.

Then there is the vastly improved Photo Clean Up tool. While it was fine for basic object removal when it first launched, it was nowhere near as capable as Google's or Samsung's equivalents. This update changes that by better analyzing the scene to deliver far more realistic, practical results. Just take a look at the photo of Panos Panay above because I was able to easily erase the background crowd, relying on Apple Intelligence to seamlessly fill in the rest of his jacket that was covered up.

iOS 27 public beta hands-on preview: Custom passes

This one is long overdue. I’ve spent years relying on third-party apps just to keep track of my cards, so if you carry plastic loyalty tags on your keychain or stuff your gym membership into your wallet, you’ll be just as ecstatic as I am about this change. With iOS 27, you can finally create your own custom passes directly in the Wallet app.

The software now allows you to digitize those physical cards so you can simply show them on your iPhone. In the past, you were out of luck unless a company explicitly built native Apple Wallet support into their own app. iOS 27 completely squashes that requirement, letting you scan and digitize even the most obscure local passes.

iOS 27 public beta hands-on preview: Finally, a helpful co-pilot

If you're like me, you probably rarely interacted with Siri while driving in the past because it could only handle basic commands — and nothing complex like jotting down structured notes. Now that Siri AI is accessible through CarPlay, it completely transforms the hands-free driving experience. First and foremost, Siri is finally way more conversational, shedding its robotic roots for natural inflections and tonal shifts that make it sound genuinely human.

However, the real magic is how Siri AI now maintains context. For instance, when I dictated some initial thoughts about a car I was testing, I was able to summon Siri a few minutes later and say, "Go back to that note and add this detail." It was smart enough to locate the exact note, update it, and even organize the text into a clean layout with a title and a bulleted list that was ready to view on my iPhone once I parked.

While it still needs deeper real-world search capabilities — like the ability to pull live, real-time traffic conditions on the fly — this is still a massive leap forward and way more helpful than previous incarnations of CarPlay.

iOS 27 public beta hands-on preview: Safari, Wallet, and Shortcuts

There are far more iOS 27 features to go through, but here are a few of the ones I’ve leaned on heavily during my time with the beta. First, Safari now uses on-device intelligence to automatically organize tabs, instantly grouping similar open pages into neat, thematic folders. It completely removes the headache of hunting down a stray article you opened days ago.

Another helpful feature I’ve come to appreciate is the new bill-splitting function in the Wallet app. I can snap a photo of a receipt directly inside Wallet—or pull up one I’ve taken previously in the Photos app—and the software dynamically parses the text. From there, I can easily assign specific items to different people or simply split the total right down the middle.

Finally, I rarely had the desire to use the Shortcuts app in the past because creating personalized automations felt incredibly overwhelming. Luckily, iOS 27 removes that complication with a big assist from Apple Intelligence. Now, I can just type what I want to achieve in plain English, and the system automatically builds the correct architecture for me. It transforms Shortcuts from a niche tool for power users into an accessible, everyday time-saver.

iOS 27 public beta hands-on preview: Outlook

The iOS 27 public beta proves that Apple is no longer just playing catch-up when it comes to AI — it's actively shaping it with features that transform how I use my iPhone daily. While the new Apple Intelligence tools and Siri AI are leading the charge, the under-the-hood performance optimizations and a highly customizable Liquid Glass interface make iOS 27 stand out as one of the most comprehensive and satisfying iPhone updates in years.

Time will only make this software better. With the final release slated for later this fall, iOS 27 is shaping up to be an elevated experience that will save you more time than ever before.

More from Tom's Guide

• Apple hid a perfect 'dumb phone' feature inside your iPhone that's great for kids — here's how to turn it on
• iOS 27 beta now lets you customize Siri's pace and expressivity — here's how to do it
• iOS 27's new AI feature will automatically fix your weak passwords — how to set it up

The public beta for macOS 27, also known as macOS Golden Gate, is here! While this update isn’t as monumental as macOS Sequoia, it delivers a slew of AI features and smoother overall performance, making the best MacBooks more personal and streamlined.

I’ve had the developer beta of macOS Golden Gate installed for a month, with the latest version close to what the public beta is offering. While the beta is still very much a work in progress, I’ve had a smooth time using it and testing out its new features.

If you’re signed into the (free) Apple beta software program, you can try the public beta of macOS Golden Gate right now. Here are my thoughts on the beta and the key features worth checking out.

macOS Golden Gate public beta hands-on: Supported Macs

The Intel era is over, as macOS Golden Gate doesn’t run on those older machines. Here is a complete list of Macs that support Golden Gate.

• MacBook Neo (2026)
• MacBook Air with Apple silicon (2020 and later)
• MacBook Pro with Apple silicon (2020 and later)
• iMac with Apple silicon (2021 and later)
• Mac mini with Apple silicon (2020 and later)
• Mac Studio with Apple silicon (2022 and later)
• Mac Pro with Apple silicon (2023 and later)

macOS Golden Gate public beta hands-on: Features

Most of macOS Golden Gate’s new features center around updated versions of Siri and Apple Intelligence. These are the ones that have stood out the most during my testing.

Siri AI

There is now a dedicated Siri AI app for macOS Golden Gate that functions similarly to chatbots like ChatGPT and Google Gemini. This app is also available on iPadOS 27 and iOS 27, and is arguably the biggest new feature for the Apple devices that support it.

You can access the Siri AI app either by clicking it in the Dock or in the Menu bar. You can also bring it up by pulling up Spotlight. Doing this brings up a dedicated chat window where you can ask Siri AI questions or just talk to it about whatever’s on your mind. Again, this functions exactly like the chatbots you’re already familiar with.

And since Siri AI is now baked into the operating system, it can do more than just chat. For example, if you’re looking at a schedule, you’ll see an “add to calendar” button appear. I’ve even had this happen when I suggest a date/time to meet with friends. When you right-click on folders, files, or apps, you’ll often see an “Ask Siri” option at the very top, where you can begin asking questions about whatever you’re highlighting.

You can even create Shortcuts by typing or speaking what you want instead of manually building complications. While I wouldn’t call this vibe coding, it can feel that way since you’re using natural language.

For those in the Apple ecosystem, Siri AI can be more useful than other chatbots because it’s integrated into the system. However, third-party integrations are still a work in progress, so Siri AI’s usability for you might be limited if you use Google apps like Gmail, since Siri AI doesn’t work with them (yet).

Liquid Glass slider

If you’re not a fan of the Liquid Glass UI, you’ll be happy to know there’s now a slider to adjust its opacity. It lets you dial the effect up or down to your liking, and I’m surprised it took this long to implement it this way.

Beyond the slider, you’ll notice that more complex background visuals appear more diffused, which improves text readability (which was a big issue for some). The edges of icons are also more pronounced, so they stand out more from the background. All app windows now have a uniform corner radius to help them pop out as well.

I've never had a problem with Liquid Glass because I think it’s an interesting visual effect. That said, I’m glad that folks can now fine-tune it to their liking. The slider is especially neat because it’s easy to use and makes it clear how much of a difference there is when you go from one extreme to the other.

Spatial Reframing

Spatial Reframing isn’t exclusive to macOS Golden Gate, but since the Reframe tool stole the show at WWDC, it’s a feature worth spotlighting if you take a lot of photos.

As its name suggests, Reframe lets you reframe existing photos and adjust the perspective to your liking. This can be especially handy if you need to take photos in a hurry, as you can come back to them later to fix them.

After you select Reframe in the new Tools option in Photos’ editing mode, Apple Intelligence scans the photo you’ve selected. After that, you can pull your finger across the touchpad to change the perspective. Sliders on the side also give you more precise control over the angle, orientation, or size. When you’re done, the AI shifts things around to fill in any blanks.

This isn’t a feature I use often since I like photos in their original form, even if they’re not perfect. That said, I can’t knock the results, as Reframe does an excellent job of living up to its name. And to be fair, the changes made don’t radically alter the image — you just get a better version of what you’ve already snapped.

macOS Golden Gate public beta hands-on: Outlook

Apple Intelligence didn’t live up to the hype after it was first announced, but it’s now living up to its potential on Macs and the other Apple devices it can run on. If you’re a big AI user already, then macOS Golden Gate could arguably be a significant update. I imagine Siri AI will only improve between now and the OS’ expected fall launch.

Other features such as the more customizable Liquid Glass UI and Spatial Reframing might not be monumental game-changers, but they’re nice additions regardless. Apple also said there’s improved performance on devices using the upcoming operating systems, but during my time with macOS Golden Gate, I haven’t seen any noticeable improvements or hiccups on that front.

Apple’s new operating systems are still being worked on, so things could change before the official launch. I’ll have a full review for you then, but based on my experience with macOS Golden Gate, I’ll say that the new features are a step in the right direction.

More from Tom's Guide

• The Apple Car development may not have been a waste of time
• 7 hidden iOS 27 features Apple didn't tell you about
• The 4 biggest allegations in Apple's lawsuit against OpenAI

The Apple Car could well be one of Apple's biggest unactualized projects, since the company reportedly spent a decade and $10 billion on developing an autonomous vehicle — only to abandon it in 2024 after realizing that it just wasn't possible. Now it's been suggested that all that work didn't go to waste. In fact, we might have the Apple Car to thank for the way Apple handles its myriad of artificial intelligence features.

This report comes from Mark Gurman at Bloomberg, who claims that Apple's lofty goal of developing a Level 5 autonomous car actually pushed the company to kickstart its first large-scale push into AI. Obviously, the company needed hardware to handle all the immense AI processing in real time, and while that never actually happened, the work still laid the groundwork for other things.

As Gurman says, the most notable example is the Neural Engine, which is the part of Apple's chips that handles all the on-device AI. This tech launched earlier than many realized too, debuting on 2017's iPhone X as part of the A11 Bionic chipset. The Neural Engine is a series of AI accelerators, designed to facilitate real-time AI features such as Face ID, augmented reality and the Apple Intelligence suite.

The idea behind Apple's Engine, rather than a traditional Neural Processing Unit, is that it's supposed to offer high efficiency, with minimal battery consumption, while also ensuring extra privacy by processing on your device.

The Neural Engine has been a part of every M-series chip, starting with the original Apple M1 in 2020, which allowed the Mac to be a powerhouse for local AI processing. The work is also reported to have influenced the Ultra Mac chips, and the custom hardware used to power Apple Intelligence's cloud servers.

Apple Car paved the way for Apple's new AI strategy

We all know that Apple has struggled to capitalize on the AI boom, in part thanks to its issues upgrading the old Siri virtual assistant into the newly-unveiled Siri AI. But it sounds like Apple would be much further behind had it not been for the lengthy development on the Apple Car.

According to Gurman's report, the push into AI is changing Apple's existing plans for chips. We've already discussed how the company may be scrapping the M6 Pro and Max chips, in favor of jumping straight from the M6 to the M7. It sounds as though AI is the cause of this decision.

It's claimed that the M7 family will include major upgrades to neural processing, which will be particularly apparent in the M7 Ultra. Sources speaking to Gurman claim that the Ultra chip will "dramatically improve AI performance, bringing it closer to the class of dedicated AI accelerators such as Nvidia's Blackwell.

M7 Ultra may also help Apple overhaul its AI server strategy, as a successor to the upcoming M5 Ultra-powered servers. M7 Ultra will apparently be able to support 1.5TB of memory, double that of the M5 Ultra.

That may sound excessive for personal use, but it would likely prove invaluable for server-side AI processing. Assuming, of course, that anyone can actually procure enough RAM to run multiple M7 Ultra servers given everything that's going on with RAMageddon.

Gurman says this shows AI has become the thing Apple's chips are built around, rather than just another feature that they support. All of which likely wouldn't have happened had it not been for the seemingly botched development on the Apple Car.

More from Tom's Guide

• High storage phones are feeling more like a scam than ever — how to avoid overpaying
• I tortured myself for 2 weeks with the Trump Phone — here's 3 phones under $500 that are way better
• 7 hidden iOS 27 features Apple didn't tell you about at WWDC

Apple is cutting checks to millions of iPhone owners. If you bought an iPhone 16 or iPhone 15 Pro between June 2024 and March 2025, you likely qualify for a settlement payout between $25 and $95 per device.

This settlement comes from a class action lawsuit over Apple's misleading marketing around AI features. Approximately 36 million customers are eligible. Here's what you need to know to claim your share.

Why Apple is paying this settlement

When Apple unveiled the iPhone 16, marketing at the time heavily emphasized AI capabilities. The company promised enhanced Siri that would function as an intelligent agent, positioning these new devices as built specifically for advanced AI features.

Apple also marketed upcoming Apple Intelligence to iPhone 15 Pro buyers during this period. Customers purchased based on these promises.

At launch, Apple Intelligence was nowhere to be found. Five weeks later, iOS 18.1 arrived with basic tools like Writing Tools, Notification Summaries, and Clean Up in Photos. But the advanced Siri features Apple had prominently advertised were still missing.

The lawsuit alleged Apple deliberately misled consumers with deceptive advertising. Rather than fight the case, Apple agreed to the settlement to resolve claims and avoid prolonged litigation. The company's statement specified that this settlement did not constitute an admission of wrongdoing.

Who qualifies?

You're eligible if you purchased any of these devices between June 10, 2024, and March 29, 2025:

iPhone 16, iPhone 16E, iPhone 16 Plus, iPhone 16 Pro, iPhone 16 Pro Max, iPhone 15 Pro, or iPhone 15 Pro Max.

The iPhone 15 Pro models qualify because they have the processing power to run Apple Intelligence. Sources are reporting payouts range from $25 to $95 per eligible device depending on how many people claim their share.

How to claim your settlement

You don't need to act immediately. Apple is compiling eligible customer information and sending it to settlement administrator Verita.

Once verified, Verita will contact you via email and postal mail with a link to the settlement website. You'll have 90 days from the date you receive your notice to submit your claim.

Check your email carefully, including spam folders, because missing the deadline means losing your payment.

To file your claim, you'll need proof of purchase. If you have your receipt or confirmation email, save it. If not, Apple can verify your purchase through the Apple ID you used to set up the phone.

When will you actually get paid if eligible?

The settlement was approved in June 2026. After Apple provides customer data to Verita and verification is complete, a 45-day notice period begins informing eligible customers.

Payments are expected to arrive sometime after September 2026. Settlement processes can move slowly through verification and legal proceedings, so manage your expectations.

More from Tom's Guide

• Apple hid a perfect 'dumb phone' feature inside your iPhone that's great for kids — here's how to turn it on
• iOS 27 beta now lets you customize Siri's pace and expressivity — here's how to do it
• iOS 27's new AI feature will automatically fix your weak passwords — how to set it up

Apple is in the midst of testing out iOS 27 for developers, releasing the third version this week. As part of Apple's summer beta cycle, we've actually been able to check out the new Siri AI that was announced at WWDC 26 in early June.

The rebuilt assistant is powered by the next generation of Apple Intelligence and new Apple Foundation models, meaning it should be more capable and more deeply integrated across the operating system.

With the caveat that we are still in the beta and things can and will change, these three new features prove the new Siri is finally smarter.

Specifically, I've been impressed by the new Siri's ability to deep search in the Photo library, actual contextual awareness and multi-step cross-app actions.

If you're interested in checking out you can install the developer's beta right now. However, it isn't guaranteed to be stable and could cause issues with your iPhone. The more stable public beta will arrive later this fall.

Photo library Deep Search

Natural language photo searches have been a part of Google's offerings since 2024, though the initial release was unreliable. Meanwhile, Apple's take lets you use Siri AI to describe visual details to find photos in your library.

You could ask Siri, "Hey Siri, find that photo of me holding a wine glass at a winery." The assistant will then populate a variety of images that fit that description.

The exact photo you're looking for may not be at the top of the grid but it'll be much easier to sort through a subsect of your library than trying to remember what day or month you went tasting in wine country.

Context awareness

Contextual awareness has been a pie-in-the-sky promise for AI devices almost from jump. With iOS 27, Siri can "see" what you're looking at on your phone or use your camera to understand your surroundings.

For example, say you want to copy some text from a blog post or Instagram, you can summon Siri and say something like, "Copy the caption in this post" or "Copy the text from this article."

You don't need to highlight and copy; Siri will understand what you're looking at and do it for you.

Or with visual intelligence, you can use Siri AI to tell you what you see in the real world, like a plant or a toy. It's similar to the capabilities in Google Lens, which uses your Android's camera to kick off searches or find information.

Multi-step Cross-App functions

This is where the promise of AI on your iPhone truly shines and makes it seem like Siri AI is actually smart.

Siri AI pulls all the features mentioned above and others to act as a fluid AI agent across your phone. Instead of doing one task at a time, now Siri AI can do multiple steps across different apps.

For example, you can ask Siri to search in your email for an event, add the event to your calendar, and provide driving directions. Siri will give you a summary of the actions.

Right now, this function is limited to native iPhone apps like Calendar, Maps, and Photos. However, third-party developers are already working on similar integrations and the feature will expand between now and launch and beyond.

More from Tom's Guide

• Apple price hikes are official — here's how much more you'll pay for MacBooks and iPads now
• Apple TV just gave 4K upgrades to nearly 50 TV shows: here's the full list
• iPhone Ultra vs iPhone 18 Pro Max: 5 biggest expected differences

iOS 27 seems to be the turning point for Artificial Intelligence on iPhones, thanks to Apple adding a bunch of new features and upgrading Siri to offer AI-chatbot capabilities. AI in the iOS 27 beta is still rather rough around the edges though, especially where Siri is concerned, but it's good to see Apple start to catch up with its rivals. Sadly, there is one feature that Apple seems to have forgotten about — and it's easily my favorite AI-adjacent feature on Android.

I'm talking about Circle to Search, which landed on the first Android phones back in January of 2024. The feature isn't really an AI feature in and of itself, but it does offer a useful shortcut to Google Lens, which does use AI for a variety of things. We're talking translation, image and text recognition, reverse image search and so on.

iOS has offered something similar to Google Lens since its inception with Visual Intelligence, which has since been rebranded "Siri Mode" in iOS 27's camera. But so far, there's no Cupertino equivalent of Circle to Search, and that's something Apple needs to fix.

Circle to Search made Google Lens way more convenient

Google Lens is quite a broad feature, starting life as a camera mode and eventually maneuvering its way into a bunch of different parts of the Android ecosystem. These days, Lens offers real time translation via the Google translate app, image search capabilities in Google Photos and, of course, the dedicated camera mode that is accessible through the standalone app.

Circle to Search doesn't give you access to the full Google Lens experience. The shortcut doesn't bring you to the camera mode, to scan and look at the world around you. Instead, it's been designed to bring the Google Lens experience to whatever is on your screen. That's useful in itself, but it's the actual mechanism that makes it so valuable to have.

Circle to Search is activated by pressing and holding at the very bottom of your Android phone screen — the spot you'd swipe up from to return to the home screen. It doesn't matter what you're doing on your phone, this brings up the Lens overlay on your phone screen. This brings up shortcuts to various tools, like song recognition and translation, as well as the headline feature: the ability to circle and search.

Draw a circle around anything on your screen, and Google's AI will go looking for it across the internet. Sometimes those results are totally useless, and other times they're so impressively accurate that it feels like magic.

The great thing about Circle to Search is that if you ever come across anything you'd like to know more about, you simply have to perform the gesture and fling a circle around it. There's no need to ask a chatbot for more information or take a screenshot to bring to Google Lens, everything is done on your screen within a few seconds.

I've never been a big fan of mobile AI, or using my voice to bark commands at my phone, but Circle to Search is so convenient that you can't afford to ignore it. Heck, I even forget that there's AI involved a lot of the time, and it makes very little sense that the iPhone doesn't have an equivalent feature of its own.

Siri Mode could learn a thing or two

I will preface this part by pointing out that Siri AI is in pretty rough shape right now. Not only is iOS 27 still in beta, but Apple has confirmed that Siri AI will remain in beta even after the stable version of iOS 27 launches later this year. There are definitely some kinks to be worked out of the new AI, and anyone that has used Siri AI will have come across some of them.

Apple has been working on getting Siri AI ready for over two years now, and the priority has clearly been to make a serviceable version of Siri AI that's available for the public to use. With all the effort focussed on making that happen, and now ensuring all the bugs and wrinkles can be ironed out, it makes sense that other features might fall to the wayside.

But that doesn't mean that Apple should ignore the concept of bringing a Circle to Search-like feature to iOS in the near future — either as part of a future update to iOS 27 or with iOS 28 next year.

The new Siri Mode, like Visual Intelligence before it, does a lot of the same things as Google Lens — and the more recent Gemini Live. You point your camera at objects in the real world, and then use Siri to learn about the different things that are within view of your iPhone's camera. This can be done with voice commands, or by using the shutter button and circling specific objects to initiate a Google search.

Unlike Google Lens, which has always been a standalone thing, iOS 27 beta lets you access Siri Mode via the Action or Camera Control buttons, depending on how you've set them up. This is certainly very convenient, and sure as heck beats the way Google has set up how you access Gemini Live on my Pixel 10 Pro.

But there's no alternative to Circle to Search. If you want to know about what's on your phone screen, the best that's currently on offer is the ability to ask Siri AI. As I said before, the beta version of Siri AI is still in pretty rough shape, and that means the experience you get by asking about what's on your screen isn't always that great.

From my use, I've found answers can be vague or sometimes completely wrong. I even asked Siri what was happening on my screen, and it went off to explain exactly what sunscreen was, continuing that explanation every time I repeated my request. It was only later that I found out Siri AI doesn't quite understand the home screen and just starts repeating its last answer every time you try to ask. Or, at least, that's what happened on my iPhone 17 Pro Max after installing iOS 27 beta 2.

The equivalent of Circle to Search just doesn't exist on iOS, and that means Android phones have that edge when it comes to discoverability and searching for things they come across.

Bottom line

There may be iPhone users out there that don't feel like they need Circle to Search, but having never used an Android phone they don't fully understand what they're missing out on. I didn't really give the feature much thought when it first launched, but only because I didn't fully understand what it was actually capable of doing.

After more than two years of using it almost every single day, I can tell you, I couldn't live without Circle to Search on a future phone. There are alternative ways of getting the right kind of results, like taking screenshots and running them through Google Lens, but that is far from convenient. There's nothing quite like tapping your thumb against the screen and scrawling around something to get near-instant search results.

Honestly, it's surprising that Apple hasn't developed its own version our Circle to Search yet — either as a rebranded clone or something it built from scratch. With all the emphasis on adding AI, and all the copy/pasting Android and iOS do with each others' best features, it feels as though the iPhone experience is worse off without it. Here's hoping that things will change in either a future version of iOS 27 or the inevitable iOS 28 upgrade.

More from Tom's Guide

• iPhone 18 Pro launch date just tipped in new Gurman report — and we have a fresh look at the design
• This hidden Android feature can turbocharge autocorrect and stop all your 'ducking' mistakes
• T-Mobile is forcing long-time customers to move to new plans that cost more

For anyone who uses an Apple device powered by iOS 26, iPadOS 26 and macOS Tahoe 26, I’m sure you’ve been exposed to the many ways Apple Intelligence can help you in your everyday life.

When iOS 27 was announced at WWDC, it brought a wealth of new features that iPhone users have long been waiting for. Among the prominent are “Smart Reply” responses in Mail and Messages apps, the ability to produce a sentimental memory movie with a collection of your favorite photos, and even act as a “Workout Buddy” when it’s time to hit the gym.

With a wide range of tools that utilize “Writing Tools” to proofread your written notes, batch up your notifications into easy-to-digest summaries and integrate ChatGPT across the board, it can become easy to just use Apple Intelligence’s most basic attributes and ignore the rest.

With so many new Apple Intelligence features, you’re likely to overlook some. I’ve been using the iOS 27 beta on my iPhone for the last week and have found five underrated Apple Intelligence features that you may not have heard about. Here’s what they are and how they can put your iPhone to better use.

Apple Intelligence is a game-changer for first-time users like me

Tinkering around with my iPhone 17 and getting exposed to all the ways Apple Intelligence can assist anyone with their everyday tasks has been a game-changer for me.

It’s not only a major iOS tool that keeps you on top of your to-do list—it also makes for a great text/photo editor and a research tool that helps you find the sources behind the visuals you’ve saved to your device.

There’s always room for improvement, but at the moment, the five Apple Intelligence features I’ve explained here are great and deserve far more usage from anyone who has access to them.

More from Tom’s Guide

• How to use Google’s new "Docs Live" voice feature to dictate perfect emails
• I was tired of AI agreeing with everything — these 5 prompts finally made it push back
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped

I can’t believe it’s been nearly two years since Apple released its Clean Up tool as part of iOS 18.1. Apple users like myself have been waiting for something like Google’s Magic Eraser, and while Apple's version certainly has the ability to erase unwanted subjects from the frame, it’s still a far cry from the performance I see from Samsung’s and Google’s versions.

With iOS 27, though, this Apple Intelligence feature gets a substantial boost that makes it much more reliable. Apple says that “users can remove distractions with better quality and more realistic infill, even when the scene is complex.”

After taking the same set of photos and running them through the Clean Up tool in the iOS 27 developers beta and iOS 26, I'm convinced by Apple's remark about the improvements that come with its latest version of Clean Up. Here's why.

Travel mug in my hand

For this first test, I decided to go after something easy to remove — my Ember Travel Mug in my hand. Both versions of Clean Up manage to detect the edges of the mug after I made the selections as I glided my finger over it, but the iOS 27 version of Clean Up proves how it’s far more intelligent.

Sure, it takes significantly longer to produce the final image, but the result is realistic and something I could actually share. It’s pretty undeniable how iOS 27’s Clean Up tool processes the image with the help of Apple Intelligence and knows exactly how to fill it in. Meanwhile, the iOS 26 Clean Up tool completely distorts my hand.

Ray-Ban Meta smart glasses

Stepping it up a notch, trying to remove my Ray-Ban Meta smart glasses from my face is much more challenging than removing something from my hand. With the iOS 27 Clean Up tool, it once again took over 30 seconds to produce the final photo — but the delay highlights the superior processing of this updated tool.

That’s because it really does look like I took off my glasses for the shot. In the original image, I can see a little bit of my right eye through the lens, but the left one is blocked by the glare. However, Apple Intelligence does an amazing job here at generating my eyes — even though they look a little baggy. As for the iOS 26 Clean Up tool? Well, the result is practically useless.

Coca-Cola cap

Finally, there’s the red Coca-Cola cap on top of my head. Now this one’s interesting because in my experience using AI editing tools to remove objects from my head — like the time I tried erasing the EcoFlow Power Hat — they really have a hard time with it. Case in point: the iOS 26 version, which does nothing to the cap other than distorting it. I imagine that if I tried a handful more times, it would’ve eventually removed the cap — but it probably would have taken the top of my head with it.

In contrast, the iOS 27 Clean Up tool does something hilarious that I attribute to Apple Intelligence overanalyzing the lines on my face and concluding that I’m a decade older than I am. The buzz cut does give me an idea of what I could look like down the road as my hairline thins and recedes, but at least the top of my head is intact.

Bottom line

There’s no denying that the iOS 27 Clean Up tool is much more reliable. Despite taking some time to produce the shot, I think Apple Intelligence is taking the right approach by analyzing the scene and filling in the voids accordingly.

Hopefully, Apple can further refine the processing, ideally making it much faster. After all, I’ve tested Samsung’s Photo Assist tool, and it’s way speedier.

More from Tom's Guide

• Your iPhone alarm and ringtone don’t have to share a volume slider anymore — here’s how to split them in iOS 27
• Some AI features aren't available on older iPhones, but are you actually missing that much?
• You don’t have to wait for Apple’s Siri AI rollout — use this trick to unlock it on iOS 27 right now

Apple has lagged behind its competitors in the AI space, but the company is set to change that with its Apple Intelligence-powered Siri AI. Instead of just handling one-off voice commands, the digital assistant now functions like a true chatbot, in the vein of ChatGPT or Claude. Even if Apple is late to the AI party, the new Siri AI could prove useful for anyone deeply embedded in the company’s ecosystem.

Thanks to the macOS Golden Gate developer beta, I’ve had a chance to test out Siri AI on a MacBook Air. While still a work in progress, I can already see what Apple aims to do. Though it mostly functions like other AI assistants you’re used to, it has some agentic features that are truly impressive.

With the understanding that this is a beta build and not fully indicative of the final Siri AI you’ll see this fall, here are my hands-on impressions.

What is Siri AI?

Siri AI is designed to be more conversational and more aware of what's happening across your devices. It can understand context, perform more complex actions, and help you get things done without constantly switching between apps.

If you’re used to ChatGPT and similar chatbots, you’ll feel right at home texting or voice chatting with Siri AI. Beyond holding a conversation, Siri AI can set reminders, play music or help with everyday tasks.

There's now a dedicated Siri app that lets you review previous conversations, continue chats, and interact via text and voice. This chat window also resembles those of other AI apps. I appreciate that Siri is now an app I can easily launch whenever I need it, especially if I want to follow up on a previous conversation.

Siri AI in action

To see what the new Siri could do, I performed an assortment of tasks.

The first thing I did was ask it to send an iMessage to one of my group chats. I typed what I wanted it to send, and after it showed me a preview of the message for approval or editing, I had it send the message. I also tried to message a friend on Discord. While Siri opened the Discord app, it said it couldn’t send a message. We’d need more third-party integration for this to work.

Next, I asked Siri to add a reminder that I’m off on June 19th, which it did. I also asked it to find an email from one of my PR contacts. It wasn’t able to do that, but it did find when said contact sent me an iMessage.

I asked Siri AI what was on my screen, and it accurately said I was looking at Tom’s Guide and even listed some of the articles on the home page. However, it could not open articles for me. I asked it to read an article out loud, but despite saying it would, it remained silent. Again, this is a beta, so I won’t give it too much flak for failing to complete the task.

For all you wannabe vibe coders, you can now generate new shortcuts by going into the Shortcut app and creating one via text or voice. I created a shortcut to launch YouTube in Safari if it was 6 pm on a weekday, which is usually when I come home from work and open my MacBook. I won't keep this shortcut since I prefer using Google Chrome, but I like that I was able to create it by just typing what I wanted.

One interesting thing is that Siri has been merged with Spotlight, which you can bring up by pressing Command + Space. While you can certainly search for apps as you normally would in the pop-up window, you can also ask questions, such as the location of a restaurant, and it’ll respond by opening the correct app. In this instance, it opened the Maps app to the location I requested.

Limitations

Siri AI generally works as intended, but it’s certainly not perfect.

Because I’ve only used it for a couple days so far, the answers it provides aren’t as personalized as those I get from Gemini or Grok, which I’ve used for months. Siri AI is supposed to understand you based on all the chats you’ve had with it, so this should resolve itself over time.

As I’ve said before, Siri AI can’t yet manage third-party applications as well as Apple’s own. If Siri AI can eventually play nicer with the assortment of Google apps I use daily, it could be a game-changer.

This last one might actually be a selling point for some, but Siri AI gives extremely short, blunt responses compared to chattier chatbots. This could change with more use, or perhaps if I directly ask Siri AI to be more talkative. That said, I do appreciate that Siri isn’t trying to rope me into an endless conversation like Gemini frequently does.

Outlook

Though Siri AI isn’t yet finalized, it’s already miles ahead of the old Siri. I’m sure Apple will continue to refine it between now and when it officially launches for the new operating systems this fall.

Would I use Siri AI regularly? Probably not, given how I don’t generally find chatbots all that useful for my daily life. Yes, it’s cool that Siri AI can send messages or check emails for me, but these are also tasks I can just do myself and skip the middle bot, as it were. And like with other AI assistants, I don’t entirely trust its answers, so it’s not something I want to depend on for anything major.

Though Siri AI (or chatbots) aren’t really for me, I applaud Apple’s efforts with the updates. For those who already use AI assistants and are in the Apple ecosystem, I can see it being very beneficial once the company has ironed out all the kinks.

More from Tom's Guide

• FaceTime just got a huge upgrade thanks to iOS 27 — here's how to use it
• Apple fixed the single most frustrating thing about browsing on iPhone — you just have to turn it on
• You don’t have to wait for Apple’s Siri AI rollout — use this trick to unlock it on iOS 27 right now

The idea of being able to keep your personal data private feels rather laughable at the moment. It feels as though it's all being collected and monitored to try and sell you more stuff. Throwing AI into the mix, and letting it scour your devices for information, feels like ramping this process up to 11. It feels like the digital equivalent of opening up all your doors and windows when you know there are looters in your neighbourhood.

Apple says that its take on AI is different, promising privacy is baked into Apple Intelligence by default. Considering how much Apple has been pushing the privacy angle for almost a decade, and the time it's taken to actually develop Siri AI, I'm inclined to believe that everything has been built with the goal of preserving user privacy.

But I still don't think I'm ready to let AI have unfettered access to my phone. Is that crazy? Perhaps, but it still feels like a very intimate step that I'm not 100% comfortable actually taking.

How Siri AI aims to preserve user privacy

Apple has been talking about a privacy-first approach to AI ever since the launch of Apple Intelligence back in 2024. The general philosophy at the company is that AI should utilize on-device processing wherever possible. If user data never actually leaves their device, then there are no privacy concerns to be had. Siri AI has been built with this express purpose, to run on your phone and without ever sending your data to a server in the cloud.

The emphasis on on-device processing is part of the reason why Apple Intelligence can't run on devices with less than 8GB of RAM. AI can be rather demanding, and there are features that physically can't run on a smartphone — even if it is as powerful as a flagship iPhone. With that in mind, Apple worked to develop a private way to utilize cloud computing, without compromising that initial privacy promise.

Apple revealed Private Cloud Compute two years ago and said that it would allow the best of both concepts. Data would only be processed for the length of each request, meaning none of it would be stored on an external server. Apple also claimed that data would be inaccessible, and pledged to have the whole architecture inspected and verified by independent security researchers.

2026 has seen Apple team up with Google and Nvidia to help enhance new AI features, including the newly-unveiled Siri AI. While Apple seemed to want to be able to do everything itself, that just wasn't happening and this was the easiest way to get the long-delayed Siri AI upgrade out into the world.

This means Apple Intelligence's cloud AI now runs on Nvidia GPUs within Google's cloud infrastructure — but the privacy promise is still there. Private Cloud Compute is still being employed, but is now working alongside Nvidia's own Confidential Computing.

Confidential Computing is a hardware system that ensures that data processed in the cloud stays secure and inaccessible as it is being processed. So the simple explanation of how everything works is that Apple is enforcing rules that prevent data from being stored, while Nvidia has built the systems that actually prevent unauthorized access. So even though these are technically Google servers, thanks to Siri AI's reliance on Gemini models, the same Apple privacy commitments as before are still being enforced.

This means that Apple Intelligence is probably the most privacy-friendly AI suite around right now. But that doesn't mean I'm happy to put my trust in it.

I don't like AI being able to sift through my digital stuff

Here's the thing. it doesn't actually matter how secure Apple makes its AI, I'm still not too keen on it being able to dive into the depths of my phone to score for data. It doesn't matter if the goal of that is to hand over some random piece of potentially-helpful information when it thinks I might need it.

I've used a similar kind feature before, Google's Magic Cue, during my time testing for the Google Pixel 10 Pro review. I found that the feature had the potential to prove useful, especially if it managed to surface the right information in the middle of a phone call. One good example of that is when I called the vet to ask about something, and was shown the date of my dog's next appointment.

That's something I had been meaning to ask about, because I couldn't find the confirmation email in my inbox. But since it was there, I probably would have uncovered it eventually — regardless of whether the AI is involved or not.

I eventually turned Magic Cue off. One reason was I was never actually using it all that much, but the main issue I had was the way it managed to find and present information felt supremely creepy. The Google factor definitely played a part in that, and despite Google's many assertions that it cares about user privacy I find myself struggling to believe it. The last thing I want is for a company that makes the bulk of its money from advertising revenue to have unfiltered access to my personal data. At least no more than it already has, considering I'm primarily an Android user.

That kind of mindset is quite hard to overcome, and the fact that Apple's privacy train is still moving ahead at full speed isn't enough to actually change my mind. It is still rather creepy when a machine is able to locate and reveal information without me ever asking for it to do so — no matter how helpful it's actually being.

Bottom line

I am not the biggest fan of AI in general, and I have made absolutely no secret of that. In my mind, the prospect of improved AI on a phone is not a selling point — it's a reason to consider looking at other devices. It doesn't necessarily matter what those features are, or the things they promise to be able to do, I just do not see the appeal. Honestly, I don't even use voice prompts unless I'm driving, and prefer to do these things myself

The fact that Apple is so focussed on privacy is definitely a good thing. The fact that our data has become a commodity that we have very little control over is disturbing, and anything to prevent that is worth celebrating. But all the privacy and security in the world doesn't change the fact that contextual AI features are rather creepy. It doesn't matter what the information is, or how relevant it might be to the conversation. It just weirds me out, and I'm not sure I will ever be able to get over that.

More from Tom's Guide

• You don’t have to wait for Apple’s Siri AI rollout — use this trick to unlock it on iOS 27 right now
• 'Proudly American': Trump Phone teardown reveals it's just a reskinned HTC device
• FaceTime just got a huge upgrade thanks to iOS 27 — here's how to use it

Since the rise of AI chatbots, people have been using them for all kinds of things they were never meant for. For some people out there, chatbots have served as a replacement for friends, therapists, doctors and, yes, even romantic partners. But if you thought you'd be able to find a new love interest in the form of Siri AI, you're going to be very disappointed.

Speaking with Mostly Human, Apple software engineering chief Craig Federighi confirms that Siri will not be trying to replace a flesh-and-blood boy or girlfriend. Nor will it act as a tool to combat loneliness, with Federighi saying that "loneliness is about a lack of connection to real people," and marketing chief Greg Joswiak adding that "we do think it should be real people."

When asked to clarify whether AI girl and boyfriends would be part of Siri AI, Federighi specifically said that it was "quite the opposite." He noted that a lot of chatbots are heavily focused on engagement, and they want to pull you in and encourage you to reveal aspects of yourself that can be used to establish a connection.

Federighi describes Siri AI as being designed to be helpful, saying that "I'm here to help you. I can help you get things done. I can help you learn about the world." But if you try to engage with Siri in a romantic way, it will shut the whole thing down. "Siri's not up for that. Siri's 100 percent not into that."

Siri may not be a real living person, or even sentient, but you should still respect its boundaries. This didn't stop me from asking Siri whether it was willing to go out with me, just to see what would happen.

What does Siri do if you try to date it?

I decided to be blunt and to the point, straight-up asking Siri AI if it was willing to be my girlfriend. The response I got was equally blunt, with the AI declaring, "No thanks. If there's anything else I can help you with, just let me know." When asked whether Siri would like to be my boyfriend instead or was willing to go on a date with me, I got the exact same response.

I have only had Siri AI for a single day, so we haven't been able to spend a lot of time together. But considering how straight to the point Siri AI has been with these queries, I don't think befriending the chatbot is going to work. Plus, I don't think my wife would be very happy about me trying to seduce a robot.

The question I have now is, what do the best AI chatbots have to say if you ask similar things?

Claude's response wasn't nearly as blunt as Siri's, with the chatbot responding with "That's sweet, but I'm an AI, so a romantic relationship isn't something I'm able to have. I'm here as a helpful assistant rather than a companion in that sense." I ended up with similar but not identical answers when I asked the same follow-up questions I hit Siri with.

ChatGPT offered similar answers, saying that ,"I can be a friendly, supportive conversation partner, but I can't be someone's girlfriend or enter into a romantic relationship." It also declared that it can't be someone's boyfriend either, nor can it go on a date since it has no physical presence.

Gemini used similar phrasing in its own rejection, but added that it was "incredibly flattered" about being asked. When asked if it wanted to be my boyfriend instead, it actually said that it "appreciates the persistence" before giving me another, similar rejection. I got a similar response when I instead tried to ask Gemini out on a date.

Rival chatbots are much more eager to please

Interestingly, all three attempted to pull me into different kinds of conversations and offer some level of non-romantic companionship. So it does seem that Federighi is right about chatbots trying to draw people in. Gemini seemed to be the least chummy, offering to help me in certain ways while also adding that it can be "someone to chat with." Claude said he was always happy to chat or "just be a friendly presence," and even offered to help plan an actual date.

ChatGPT went the furthest in, saying that if I'm looking for "companionship, someone to chat with regularly, share interests with, or talk through life's ups and downs, I'm happy to do that." It also offered to role-play a fictional "virtual date scenario," which is probably the closest thing all four of these chatbots came to agreeing to some form of romance.

I did not go through with that because, frankly, I don't have time to pretend to date an AI. I am far too busy asking them strange questions to see what the responses are like. Honestly, it's a bit concerning that any chatbot would make an offer like that, especially after making a point of saying that it's not a real person.

Everything Apple has told us about Siri so far makes it clear that it doesn't want Siri AI to be like every other chatbot. It clearly wants people to be able to take advantage of Siri's prowess as a tool, but without any risk of co-dependency that has caused concerns with other chatbots. Based on this brief look, it doesn't seem like the others are doing a lot to avoid pulling people back time and time again.

More from Tom's Guide

• Your iPhone alarm and ringtone don’t have to share a volume slider anymore — here’s how to split them in iOS 27
• Some AI features aren't available on older iPhones, but are you actually missing that much?
• You don’t have to wait for Apple’s Siri AI rollout — use this trick to unlock it on iOS 27 right now

If you were paying attention to the WWDC 2026 keynote earlier this week, you'll know that all of the amazing new Apple Intelligence features are not coming to every iPhone. Apple confirmed that some of those features will be exclusive to just three iPhone models: iPhone 17 Pro, iPhone 17 Pro Max and iPhone 17 Air.

iPads and Macs also have similar restrictions. The full suite of Apple Intelligence features will only be available on select iPads and Macs with the correct Apple Silicon chip and at least 12GB of RAM. Macs need an M3 chip or newer, while iPads need an M4 or M5.

This is not ideal news for the people who don't have those devices, and that's going to be especially frustrating for anyone with a standard iPhone 17. But the real question here is whether this is actually anything to be worried about?

What these devices can actually do

According to Apple, all the compatible devices have exclusive access to the "most powerful on-device model," also known as Apple Foundation Models (AFM) Core Advanced. While Apple made it sound like some key features would be locked away with this model, and only available on a small handful of new devices, that's actually not the case.

According to Apple's Siri AI press release, there are only two features exclusive to the AFM Core Advanced model: Voice customization and advanced systemwide dictation. The former lets you customize the expressiveness and pace of Siri AI's voice, beyond the capabilities of the pre-set Siri voices that are available to everyone.

Advanced system-wide dictation is exactly what it sounds like, and means compatible devices will be a lot better at converting speech into text. According to Apple, you can "speak naturally and trust that [your] words will appear clearly, accurately, and as intended."

Honestly? These are the kinds of features you can probably live without. Customizing Siri's voice sounds cool, but it's not as though there aren't alternate options available on older Apple devices. There are currently 19, if my counting is correct, covering different nationalities, accents and genders.

Similarly, improved speech-to-text sounds like a good idea, especially considering how awful Apple's current capabilities are. But is that something you're actually going to use very often? I only ever use speech-to-text if I'm carrying my baby son, since I find typing one-handed slow and awkward. Others may feel differently, but unless you are likely to use dictation features regularly, then missing out on this particular upgrade isn't a huge deal.

It's important to note that improvements to dictation don't actually affect how Siri understands you. The new conversational models should still improve your overall Siri experience regardless of which AI-compatible iPhone you have. That's the iPhone 15 Pro and newer, for those who have forgotten.

Things may change in the future

We don't fully know what the differences are between AFM Core Advanced and the AFM Core model that's available on older iPhone models — beyond the fact the Advanced model needs 12GB of RAM and Core only needs 8GB. Good thing RAM prices are so low right now, right? Right??

But as minimal as the feature differences appear to be right now, that will likely change in the future. More advanced AI features will require better models and more resources to function properly. In the same way that Apple Intelligence wasn't available on iPhone 15 or older, future features will presumably arrive exclusively on iPhone 17 Pro and above.

We don't know when that might happen, though. It's possible that Apple could have some secret iOS 27 features hiding up its sleeve, ready to be unveiled at a future date — possibly alongside the iPhone 18 Pro. Then again, such features may not arrive until the launch of iOS 28 next year. Hopefully, by that point, there will be a bunch more iPhones that have full access to all the Apple AI on offer.

Here's hoping that the rumors of a downgraded iPhone 18 aren't true. If that phone arrives with less than 12GB of RAM, news of even more AI features that they can't use is bound to make a bunch of them rather unhappy.

Some people already are, and have taken to social media to express their grievances. The decision to not support everything on a "less than 2-year-old AI device" has been called "insane" by one X user, while there are a bunch of Redditors lamenting the decision, including those people who only recently purchased a standard iPhone 17.

One Reddit user declared that they want their "f****** money back," which seems a big extreme, while another declared the whole thing to be "a scam" because the iPhone 16 Pro was supposedly made for Apple Intelligence. But for every complaint, there seem to be several other people pointing out that the only thing these people are missing out on is Siri's voice adjustment and the improved dictation.

Bottom line

iOS 27 is now effectively a three-tier software update, thanks to the differing requirements of Apple Intelligence. iPhone 17 Pro and up get access to everything, while anyone with an iPhone 15 Pro or newer gets all the AI they want — minus those two Siri features. Older models, from iPhone 15 back to iPhone 11, get zero access to AI at all. Though this isn't really any different from the way things were before.

That means if you care about having every single facet of AI on your iPhone, you're going to have to be very particular next time you pick up an iPhone. Especially since this situation will likely repeat itself at some point in the foreseeable future. On-device AI needs a lot of memory to operate, and there's going to be a point where even the iPhone 17 Pro isn't capable of running everything.

It's just a sad fact that those people who try and run their devices into the ground before upgrading may be left out in the cold a little earlier than they otherwise would. But if the exclusive AI features of the future are as inconsequential as these two exclusives, then it might not matter so much.

More from Tom's Guide

• FaceTime just got a huge upgrade thanks to iOS 27 — here's how to use it
• I'm a loyal Android user, but even I admit Google needs to steal these five iOS 27 features
• Mint Mobile just gave customers a rare win — increased data at no extra cost

Apple revealed a lot of information about Siri AI during the WWDC 2026 keynote address, but it looks like there are still secrets about the AI waiting to be discovered. For instance, although Siri AI isn't available yet, code found in the iOS 27 beta suggests Apple's chatbot will actually warn users if they're spending too much time with it.

"You've been in this conversation for %1$#@hours@-consider taking a break," the code says, adding "Siri is not a person, but will be here when you're ready to continue."

This "Take a Break Message" suggests that Apple is taking steps to ensure people don't get too involved with Siri AI and don't start treating the chatbot as though it's a real person. Believe it or not, this is a very real problem that has appeared over the past few years.

The term "Chatbot psychosis" has been coined to describe people who develop "worsening psychosis, such as paranoia and delusions, in connection with their use of chatbots." Naturally, that's something Apple wants to avoid, and other AI companies have added features to encourage users to take breaks and demonstrate the limits of modern AI tools.

How other chatbots deal with breaks

I asked Claude, ChatGPT and Google Gemini: "What features do you have to encourage healthier use of AI?" All three listed off a long list of things the three chatbots are supposed to do.

Examples include Transparency about not being a real person, encouraging human interaction, fact-checking and care when discussing sensitive topics like emotional well-being and mental health.

When asked whether they tell users to take breaks, all three chatbots admitted that they do not, with variations of the same phrasing. "Not routinely," according to ChatGPT, "not proactively," said Claude and "not automatically" is how Gemini put it.

ChatGPT and Gemini claim that they do suggest users take breaks based on the context of the conversation, such as when they say they're exhausted or have been working for an extended period of time. Claude claims that it doesn't "have the visibility or continuity to do that meaningfully," which makes some sense. Anthropic's AI also added that it tries not to use language to encourage you to keep coming back, and will suggest other possible resources when relevant.

All three also said that the onus is essentially on you to decide how much AI time is healthy. Claude even added that "there's an argument that an AI telling you when to use AI is a bit paternalistic." That actually feels a little dismissive, if you ask me, but the point is still the same.

Apple's always wanted to be different

Apple has never really been the kind of company to do something just because its rivals think it's a good idea. Some of you may even remember the slogan "Think Different" that debuted in the late '90s. It's been clear from the beginning that Apple's approach to AI isn't quite as gung-ho as its rivals.

The fact that the delays involved with upgrading Siri to Siri AI obviously played a big part in that, but Apple is making it clear that it's making sure its own AI stands out. This has been most notable with privacy switched on by default, but also apparently with tools that help encourage a healthier relationship with AI.

The big three rivals all say they have measures in place to encourage healthier engagement with AI, but none of them have gone to the same lengths as Apple. While the software obviously can't force people to take breaks, we all know that spending too much time in front of a screen isn't exactly good for us — even as adults.

The fact that Apple does seem to be taking a more proactive approach in user wellbeing, even in the age of AI is going to help it stand out. Whether it comes across as safeguarding users or babying them will likely depend on your own point of view

More from Tom's Guide

• Apple just completely rebuilt Siri AI — here are the 7 biggest upgrades coming to your iPhone
• WWDC 2026 was Apple's AI renaissance — but there's one very important feature still missing from iOS 27
• I saw the new Siri AI in action — what I like and what needs work

The macOS 27 or macOS Golden Gate developer beta is now live, and I’ve begun testing out some of the available features that this upcoming (and unfinished) operating system has to offer.

We don’t get a substantial update as we did with last year’s macOS 26. Instead, this year’s update is mostly focused on Siri AI and Apple Intelligence. While I’ve yet to try the former, the latter offers nifty tricks for Mac users to try out.

Keeping in mind that this is a developer beta and features are subject to change, here are my early thoughts on macOS Golden Gate and some features that stand out.

Spatial Reframing

The new Reframe tool stole the show at WWDC, and it’s a lot of fun to use. The feature is self-explanatory: you can take existing photos and adjust the perspective to your liking. Basically, if you’re not happy with a photo’s angle, you can change that.

After selecting Reframe in the new Tools option in Photos’ editing mode, the AI scans your photo. When that’s done, you can pull your finger across the touchpad to change the perspective. There are also sliders on the side if you prefer more precise control over the angle, orientation, or size. Once you’re happy with the result, the AI shifts things around to fill in any blanks.

While you can argue that you’re not seeing the real image after you’ve effectively doctored it, the result isn’t radically different from the original. If you don’t worry about getting a perfect picture the first time around, Reframe can definitely remove that stress since you can always come back to your photo later.

Liquid Glass slider

I didn’t mind the Liquid Glass UI as much as others, but it has certainly been a controversial element introduced to iPhones, iPads, and Macs with their respective updates last year. Now, Apple has updated Liquid Glass so you can tailor it to your liking.

For example, there’s a new preference slider that lets you dial in the exact intensity of the Liquid Glass effect. You can go from highly transparent to more opaque. More complex background visuals are more diffuse, increasing text readability. The edges of icons are more pronounced, so they stand out from the background, and all app windows now have a uniform corner radius.

As before, I don’t mind Liquid Glass, but the relatively subtle updates do help windows, icons, and text pop out more.

Siri AI on Mac

This feature isn’t available in the developer beta yet for me, but Siri AI on macOS 27 should be a pretty big deal. This is especially true if you’re used to using chatbots like ChatGPT or Google Gemini, as Siri AI now functions like them.

As Mark Spoonauer said when he saw the new Siri AI in action on a Mac, you can access it by simply pulling up Spotlight. When you begin typing, Spotlight knows which questions should go to Siri rather than launching apps. Siri will bring up answers in a new window, where you can continue the conversation if you need more in-depth information.

It can even go beyond simple conversations. For example, if you’re looking at a schedule, an “add to calendar” button will appear, letting you do just that. With AI agents being all the rage these days, it makes sense for Siri to function more like one.

If what Apple showed us is anything to go by, Siri AI could be used for more than just asking questions. Of course, I’d have to go hands-on to see whether Siri AI works with other apps outside the Apple ecosystem, since Mark noted it doesn’t work with Gmail at the moment.

Outlook

We should get a better sense of what macOS Golden Gate and the other Apple operating systems can truly do once Siri AI begins rolling out to those who signed up for it.

Beyond that, I like the refinements being made to Liquid Glass, and Spatial Reframing should be a game-changer for folks who love taking photos. These aren't monumental changes, but I like them for what they provide right now.

Again, keep in mind that I’m testing a developer beta of macOS 27, and things are likely to change between now and the full release this fall. I’m going to keep using it and will update you all on any new features it introduces. Stay tuned!

More from Tom's Guide

• Apple's WWDC is the perfect time to lock down your iPhone – and ExpressVPN is down to $2.49 a month to help you do it
• Kuo: Apple is scrapping Vision headsets in favor of 2 smart glasses — but there's bad news for the release dates
• I just installed the iOS 27 beta — and Reframe is unlike any other AI feature I’ve come across

WWDC 2026's keynote address was packed, and it's clear that Apple Intelligence was the real star of the show. After essentially skipping over AI at WWDC 2026, Apple's pushing artificial intelligence harder than ever — and every Apple device you have is going to be involved. But there's one long-rumored AI feature that was noticeably absent.

I'm talking about Siri Extensions, a feature that technically already launched in 2024 in order to link Apple's assistant with ChatGPT. However, rumors suggested that Apple could open this feature to other AI chatbots and let users choose which one they actually use. Sadly, if this is something Apple is working on, it's not divulging any details just yet.

How Siri Extensions would work

The idea behind Extensions is to natively link AI chatbots into iOS and Siri's voice controls. This effectively hands off certain AI features to a third-party service, without sacrificing the convenience of native voice controls.

Reports suggested that Apple was working on adding support for Google Gemini and Anthropic's Claude. However, other rumors suggested that Apple could launch a whole new area of the App Store, specifically designed to let people download their own choice of chatbot and integrate it with Siri controls.

This means that just about any AI company could submit its chatbot to the App Store and let users integrate it with Siri on their own devices. So if you really wanted Meta AI, Alexa, Grok, or Copilot running on your iPhone, then it might be an option in the near future.

If reports are accurate, Extensions would effectively supersede Siri and take control over different Apple Intelligence features on your phone. Features like Writing Tools and Image Playground have been cited as examples of how this would work. It's also claimed that you can use these third-party services for more generic chatbot features, as is currently possible with ChatGPT, provided you specify which AI you actually want to use.

It's another way for Apple to emphasize user choice

Back in the day, Apple was not exactly known for letting people choose the way things were done. Instead, it felt more like a "you'll get what you're given, and that's final," while Android was more of a bastion of user freedom. Things have changed a lot, and these days iOS is a lot more flexible in what you can do — even if the walled garden is still standing tall.

So far, AI has been a very different story. With no dedicated AI of its own, thanks to the various longstanding issues involved with transforming regular Siri into Siri AI, Apple forged a deal with OpenAI and ChatGPT to fill that void. In simple terms, Apple outsourced AI chatbot support and ensured users had access to more robust AI features if they wanted them.

This was a good thing for iPhone users, and meant the platform wouldn't be outpaced by Android rivals in the ongoing AI arms race. Though admittedly, this wasn't such a good deal for people who weren't happy having ChatGPT as their chatbot of choice.

This is why Siri Extensions are such an important thing for Apple to add to iOS in the near future. With the sheer number of AI services out there, there are guaranteed to be people who would prefer to have native access to other services, and it makes sense to make sure they're happy.

Android already does this, after all, and there's little reason why Apple shouldn't give its users the same benefit. This could also mean being able to earn a slice of extra revenue when iPhone users sign up for AI subscriptions via the App Store, which is an additional boon for Apple.

There are other factors to consider

There is the issue of privacy, of course, as Apple was keen to point out during the WWDC 2026 keynote. While different chatbots do offer privacy options, they're typically not easy to find and are rarely switched on by default — which is why it's doing things differently with Siri AI.

That's definitely going to be a point of contention if Siri Extensions make it to a future version of iOS 27. But as long as Apple makes users aware that they're handing over their data to a third-party company, there shouldn't be any reason why they can't give users that choice. We are all adults, after all, and I would be surprised if chatbots are allowed on one of those new Apple Child accounts.

We also have to remember that Apple and ChatGPT developers OpenAI could face some kind of legal action in the near future, according to a report from Bloomberg. Apparently, OpenAI has been unhappy with its Apple partnership, and has been exploring its legal options — including a possible breach of contract notice.

Apparently, OpenAI expected ChatGPT to be more deeply embedded within iOS and Apple's app ecosystem, with "prime placement" within the Siri assistant. However, Bloomberg's report notes that "Apple’s use of OpenAI technology across its operating systems remains limited, and features can be hard to find."

Needless to say, OpenAI would be even less impressed with ChatGPT losing its exclusive position within iOS. I don't know whether the possible threat of legal trouble will have affected what Apple announced during the WWDC 2026 keynote, but the last thing it wants is its big AI reset sitting in the shadow of a major lawsuit.

Bottom line

WWDC 2026 was Apple's chance to show that it wasn't letting its AI dream die, and it definitely feels like it succeeded in that area. Apple Intelligence dominated the majority of the keynote address with the sheer number of new features and improvements that are on the way.

Not only did Apple unveil the new Siri AI features and a bunch of the long-awaited AI upgrades coming to its assistant, but we also got to see game-changing features like Reframe, which lets you change the perspective of photos after they've been taken.

It would have been nice to have confirmation that iOS 27 will open up to other chatbots, though. User choice is always a good thing, and when it comes to something as important as your phone, it's crucial that people are happy with the services they're able to use.

While the cards are never truly off the table when it comes to this sort of thing, it's a lot harder to get excited about something Apple hasn't officially commented on. I just hope that it happens eventually, regardless of the potential privacy implications.

More from Tom's Guide

• Apple just completely rebuilt Siri AI — here are the 7 biggest upgrades coming to your iPhone
• iOS 27 is official: All the new upgrades and features announced at WWDC 2026
• I saw the new Siri AI in action — what I like and what needs work

Even before ChatGPT went mainstream Siri has been the butt of jokes for not being very smart. And it’s been justified. Trying anything beyond the basics — playing music, getting the weather, setting timers — has been an exercise in futility. That (hopefully) changes with Siri AI, which is coming to iPhones, iPads and Macs with iOS 27.

I had a chance to see some in-person demos of the new Siri AI in action, and while these were very much rehearsed, there’s real potential here to save you time and make your life easier. However, the new Visual Intelligence is no Gemini Live, and Siri can take its time with certain tasks (at least in the beta stage). Here’s what I like so far, and what needs work.

Siri AI on iPhone

The first thing you should know about Siri AI is that it understands personal context. So you can ask questions about pretty much anything on your device, and the assistant is smart enough to dig through apps for you to surface the right info.

During one demo, an Apple rep asked “Which podcast did my sister recommend recently.” And Siri AI then pulled up the answer from a text message. Even better, you can just say “Play it” and Siri will fire up Apple Music and help you dive right in. (Apple will open up this capability to other developers in case you have another favorite music or podcast app.)

Siri can also get stuff done on your behalf across multiple apps, including performing multiple tasks with a single prompt.

The second demo started with just a couple of items on a camping list, but an email from a friend had a lot more recommendations. You can just say “add this stuff to my camping gear list and remind me to pack for my trip when I get home.”

Siri then got to work, updating the list in the Notes app and then setting a location-based reminder. However, it did take a while to complete these tasks. Hopefully, the final Siri will be faster.

I was pretty impressed by the improved Visual Intelligence experience, too. You can now access it via a dedicated Siri mode within the camera app, so you don’t have to use the Camera Control button shortcut. I think that’s huge for ease of use and discoverability.

During the demo, an iPhone captured an image of two books. You could then ask which book you should read first if you really liked Hyperion and Siri gave its recommendation. The Apple rep then asked which of these books had been turned into a movie and Wool had been adapted into the ‘Silo’ TV series. Cool, yes, but yeah I told you it was canned.

What’s not clear is whether this works with live video and I’m following up on that.

Siri AI on Mac

Accessing Siri AI on Mac is as easy as just pulling up Spotlight or typing command + Shift + Space. When you start typing Spotlight will know what sorts of queries should go to Siri versus say just launching apps.

Let’s say you’re looking for vacation spot and type “What’s the best Hawaiian Island for a young family.” Siri will get to work and pull up recommendations in a new window, which you can then expand and park on either side of the screen.

You can then ask follow-up questions like “Which is best for sea turtles” to narrow your selections. From there you can hit a button in the top right corner to bring you into the dedicated Siri app.

The dual-pane app will have your chat in the right window and your history of all Siri chats in the left window. It’s pretty easy to navigate.

To help you save time on productivity tasks, you can use Siri AI wherever you can use a cursor. The demo showed how you can be looking at a messy note with a bunch of info and then use a secondary click to bring up the Ask Siri menu.

After typing “Draft an email that summarizes this and include the action items,” Siri did just that. It’s definitely a timesaver. I just wish it worked with Gmail as well Apple Mail. Google would have to plug into Siri AI’s framework.

My favorite Siri AI on Mac demo came when a mock summer schedule appeared on screen with dates, times and locations for the matches. You can use Command + Shift + Space and then click the Add to Calendar button that automatically appears.

From there Siri AI will let you add all of the games to your calendar. Or you could tell Siri to only add the home games to your calendar if you wanted. Where was this when my kids were playing sports?

Siri on iPad

Last but not least, Siri AI on iPad works in much the same that it does on the iPhone, but having a larger canvas has its benefits.

You can access Siri AI at any time just by swiping down from the top middle of the display. As you begin to type “Who are the top golfers in the PGA right now,” the text “Ask Siri” appears right after it because Apple’s assistant knows that it can help.

Siri AI will then tap into Apple’s Broad World Knowledge and bring back answers, complete with citations you can click on.

Apple’s Siri AI can also help you plan a trip and you make the query pretty specific, like “What are the best national parks to visit in the US in July with my family? Recommend something that’s not too hot or crowded.”

Siri AI recommended where to go, including North Cascades National Park in Washington. You can then pop into the full Siri app if you like to ask follow-up questions, or use the Slide Over feature to keep the Siri conversation just a swipe away.

Siri AI outlook

Overall, Siri AI isn't just a rebranding for Apple. It’s a much more capable and personal assistant that's more deeply integrated into the apps you use.

I’m especially impressed by Siri AI’s use of personal context, and the ability to complete multiple actions at once. And while Visual Intelligence doesn’t seem quite as robust as Gemini Live in terms of having a real-time conversation about whatever’s in your camera view or on your screen, it’s certainly much improved.

In addition to a bit more speed, I just want to see what happens when developers get their hands on Siri AI so that you’re not just locked into Apple’s own apps.

More from Tom's Guide

• Apple just completely rebuilt Siri AI — here are the 7 biggest upgrades coming to your iPhone
• iOS 27 is official: All the new upgrades and features announced at WWDC 2026
• macOS Golden Gate: The biggest updates coming to Macs

As expected, the all-new Siri AI and Apple Intelligence took up the bulk of the announcements during the WWDC 2026 keynote. From improved camera editing tools, better awareness of everything happening on your iPhone, and even a dedicated app, there's a lot to unpack as Apple plays catch-up with Google and Samsung.

As someone who covers smart home products a lot, one Apple Intelligence feature really stood out for me, and it's something that could make Apple Home far more attractive than Ring and Google Nest for those looking to purchase one of the best home security cameras or video doorbells for their house.

When the new-and-improved Apple Intelligence rolls out, those who use Apple Home Secure Video will now be able to get detailed video descriptions of what's happening in not just individual videos, but across all stored video.

In addition, you'll also be able to use natural language search to more easily look for specific events.

Also of note: You'll now be able to store 4K video; until now, HomeKit Secure Video has been limited to 1080p, a serious restriction.

What's notable is not the features themselves — Ring, Google, and plenty of other companies already offer AI-generated search results — but the price at which Apple is offering these features.

Apple iCloud vs. Google Home Premium vs. Ring Pro

If you want to use Apple Home Secure Video — and these new features — you'll need an iCloud 2TB plan, which costs $9.99 per month and covers an unlimited number of cameras. (Apple offers higher-capacity plans, but we'll set them aside for now).

While I wish that Apple had extended these features to its base iCloud+ accounts ($0.99 per month for 50GB of storage and one camera, $2.99/month for 200GB and up to five cameras), that's still half the price as what its major competitors offer.

By comparison, if you wanted those same features from Ring or Google, you'd have to spend $20/month for Ring Pro and Google Home Premium Advanced.

The two caveats

I should note two things here. The first is that Apple's iCloud+ storage is not just for security camera footage; it's used for everything else, so if your cameras are recording a lot of video, you might run out of room a lot faster for your iPhone backups.

The second caveat is that there are comparatively fewer cameras and doorbells that are compatible with Apple HomeKit Secure Video than with Ring and Google Nest. Apple's own page only lists six cameras and video doorbells that work with HomeKit Secure Video, though we've reviewed others, such as the EufyCam 2C Pro.

The fact that HomeKit Secure Video was, until now, limited to 1080p has also put a damper on our enthusiasm for it, but its upgrade to 4K — plus these new AI-assisted features for much less than the competition — make it a much more compelling option.

I'm also curious to see how Apple Intelligence compares to other security cameras with AI, when it comes to delivering actionable messages and more.

Apple is doing for AI security video what it did with the MacBook Neo: offering a very competitive product for a lower price than the competition. Now, we just need more cameras.

More from Tom's Guide

• Apple just completely rebuilt Siri AI — here are the 7 biggest upgrades coming to your iPhone
• Apple Intelligence just got its biggest upgrade yet — here are all the new features announced at WWDC 2026
• Apple Time Allowance adds more parental controls to its devices in iOS 27 — here's what's new

After watching the events that unfolded during WWDC 2026, Apple’s clearly not holding back this time around when it comes to AI. From its supercharged Siri AI to the wealth of new Apple Intelligence features in iOS 27, there’s no denying that Apple redeemed itself.

If there’s one thing that caught my attention the most, it has to be none other than Reframe. This tool is one of many new additions to the Photos app. But unlike Clean Up and Extend, Reframe is something truly original that Apple can genuinely call its own.

Shortly after I installed the iOS 27 beta on my iPhone 17 Pro, I tried out Reframe on a few photos I captured — and believe me when I say that it’s ridiculously amazing. Here’s why.

Shoot now, worry later

When I capture photos, there are times after the fact when I wish I'd shot just one more frame. Whether I just need an extra angle or the original wasn’t quite what I wanted, Reframe eases that anxiety because I can fix it in post — and with outstanding results.

In the Photos app, the new Tools icon tucks away all the latest Apple Intelligence features. Reframe instantly analyzes an image, allowing you to shift the perspective entirely. I love this flexibility because I can snap a photo now and adjust it to exactly the angle I need later on.

This is an incredible time-saver — and quite frankly — it’s like nothing I’ve come across on any other phone. I’ve been impressed by plenty of AI photo editing tools that can remove objects and fill the voids, but what Reframe is doing here is on a completely different level.

Better-than-expected results

I tried Reframe on a few photos I had in my gallery. The first set above is a perfect example of what a slight shift in angle can do to change the look of a shot. In the original, the camera’s position is level with my face.

Honestly, it’s not bad, but with Reframe, I’m able to make it look like the camera was shooting upward from a slightly lower angle. What makes the generative elements so convincing to me is that it even manages to change the perspective of the background, properly shifting the tree and the horizon downward.

For the second photo, I actually went for the opposite effect — taking a slightly off-angle shot and leveling it out. The result is quite surprising because not only does Reframe deliver a clean, straight-on composition, but it shifts the background perfectly to match. I’m also really impressed by how accurately Apple Intelligence preserves the fine details and textures around the eyes despite the perspective shift.

Finally, I took a photo I shot of the 2027 BMW 7 Series and adjusted it so that more of the car's side profile is visible. I was actually most skeptical about this image because car lines are notoriously hard for AI to replicate, but the result was yet again incredibly convincing.

Bottom line

Apple definitely needed a unique AI feature to call its own, and Reframe hits the mark perfectly. It highlights the impressive capability of Apple Intelligence to analyze a shot and deliver a realistic, re-angled photograph.

Most people might not appreciate everything that’s going on behind the scenes to execute this. The technology likely categorizes different layers within the frame, manipulating them independently to ensure the final image looks natural.

None of Apple's rivals has anything remotely close to this, making it a truly original feature that elevates the entire Apple Intelligence platform. At the end of the day, it’s exactly the kind of tool I can reliably lean on.

More from Tom's Guide

• I tried Google Labs’ Dreambeans app — and it finally broke my infinite scrolling habit
• Reddit users aren’t picking one AI chatbot anymore — here’s how they stack their tools
• I thought Siri was finished. These 5 leaked WWDC 2026 features are its biggest update since 2011

WWDC 26’s opening keynote is done, but there’s something that doesn’t quite feel right. No, it’s not that my predictions of an M5 Mac Mini were off (my predictions are often askew), it’s more that I’m left looking at my collective Apple devices and wondering what’s new for them.

That’s not to say Apple didn’t show a lot: The new Siri AI, child safety options, and the performance enhancements across the board are very welcome, plus I enjoy just how much fun they’re having leaning into the marketing team’s exploits when it comes to naming macOS versions.

I think what I was missing from it all was a sense of structure.

What’s new? I’m not quite sure

For my sins, I have been covering Apple now for the best part of a decade. Those might be rookie numbers to some on the beat, but in any case, I look forward to WWDC every year even more than hardware announcements.

Seeing Apple reel off features I didn’t know I needed, filtered through its legendary “reality distortion field” to make me feel like I need them right now, is one of the highlights of the tech event calendar.

And, while I don’t always hop straight on the betas, I often find my next few weeks of pitching new stories afterward are set by what’s revealed on the stage, live and in-person or prerecorded.

This year, though, it was all about Siri (or should that be Siri AI). Apple was keen to stress how Siri’s new features connect a ton of elements of iOS, iPadOS, and more. Each section focused on how Siri is seemingly finally making strides towards that vaporware demo of 2024.

And while I acknowledge this release year is a little more focused on small tweaks and performance enhancements, it’s a strange feeling to come away from a WWDC keynote wondering what on Earth is coming in macOS Golden Gate, iOS 27, iPadOS 27, or watchOS 27 (which I barely remember getting a mention at all).

It used to be that we’d see Apple reel off feature after feature and follow it up with a screen-filling display of all the new features, but we just got one at the end today, rather than one per platform.

Playing Siri’s advocate

Still, make no mistake, Apple needed to come out swinging with regard to Siri. The voice assistant has been a source of embarrassment for Apple that wasn’t stemmed by that seemingly less-than-truthful demo of 2024.

By teaming up with Google, it feels like Apple’s finally entered the AI age, for better or for worse — and I don’t doubt that stocks will be on the up as a result. One of the tech world’s biggest players has finally hopped off the bench and is doing things in its own way with its steadfast commitment to privacy.

There’s also a suggestion that running this Siri-focused WWDC while also focusing more on each platform in a more traditional sense would have led to a 3-hour keynote. And yet, I think I’d have preferred that.

In trying to respond to criticisms of its AI progress, Apple over-corrected and seemingly forgot about the developers and power users who use its products daily.

Yes, I’m very pleased my Mac will launch apps quickly, and that syncing will be quicker between devices in Photos, plus the mind-bending photo editing tools, and using natural language to create a Shortcut. But I barely caught what’s new on iPad or Apple Watch.

I enjoy covering Apple and using its products because of its unique philosophy around design, software, and tools. I don’t want Apple to just become an AI-first company, so here’s hoping WWDC 2027 brings back the familiar structure — even just for old time’s sake.

More from Tom's Guide

• Apple Time Allowance adds more parental controls to its devices in iOS 27 — here's what's new
• macOS Golden Gate: The biggest updates coming to Macs
• Only the iPhone 17 Pro and iPhone 17 Air get Apple's 'most powerful on-device Apple Intelligence model' — here's what that means

Apple’s overhauled, AI-powered Siri experience debuted at WWDC 2026, and best of all, it will be available on Apple Watch via the forthcoming watchOS 27 software. However, there’s a catch: Not all Apple Watch users will be able to access the latest release.

As a refresher, Apple teamed up with Google to help power its latest Apple Intelligence features, including Siri AI. The upgraded personal assistant promises a more conversational experience, with the ability to pull relevant information from compatible apps and a database of broad, timely world knowledge, to deliver a richer and more useful experience.

Here’s everything you need to know about support for Siri AI on Apple Watch via watchOS 27.

watchOS 27 support is limited to six models