A 4GB file showed up on my computer — and I don’t remember downloading it.

After digging around, I traced it back to Google Chrome. According to a new report from Cybernews, some users are finding the same file, called weights.bin, hidden inside Chrome’s folders — and it may be tied to upcoming AI features.

So, the obvious question is why is Chrome downloading something this big — and did you agree to it? Here's what's actually going on.

What is the 4GB file in Chrome?

At the heart of the controversy is a file called weights.bin, which researchers say is part of an on-device AI model used by Google Chrome.

Specifically, it appears linked to Gemini Nano, Google’s lightweight AI model designed to run locally on your device instead of in the cloud. This model powers features like writing assistance, text summarization, scam and phishing detection as well as AI-powered autofill and suggestions.

So, instead of sending your data to a server, the model runs directly on your computer, which is technically better for privacy. Of course, there's an obvious tradeoff (storage).

Why is the AI model so big?

The thing is, even “lightweight” versions like Gemini Nano can take up gigabytes of storage because they include massive sets of trained parameters, hence a file like weights.bin.

This explains why some users are noticing a sudden 4GB increase in storage usage, new folders inside Chrome directories (like OptGuideOnDeviceModel) and no obvious notification explaining what was downloaded. But all of that could be a big problem for anyone using a laptop with limited space.

Did Chrome install this file without permission?

Well, this is where things get a bit...murky. According to the Cybernews report, the model may download automatically when certain AI features are enabled, but users aren’t always clearly told that a multi-gigabyte file is part of that process.

There’s no strong evidence that Chrome is installing this on every device without consent. But there is a growing concern that the download isn’t always clearly explained or that it may be tied to features users didn’t realize were enabled. In other words, there's no obvious "yes/no" moment for the storage impact. The company just wasn't exactly transparent.

How to check if it’s on your PC

If you’re curious (or low on storage), here’s how to look:

• Open File Explorer (Windows) or Finder (Mac)
• Navigate to your Chrome data folders
• Look for directories with names like: OptGuideOnDeviceModel or files named weights.bin

These locations can vary depending on your system, but they’re typically buried in Chrome’s application data folders.

Can you delete it?

Yes, you can delete it, but there's a catch. If you delete the file, Chrome may re-download it later if the related AI features are still active. You may notice that some AI-powered features inside the browser may stop working.

To fully prevent it from coming back, you may need to disable certain AI features in Chrome settings or, turn off experimental flags tied to on-device AI. Keep in mind that these settings aren't always easy to find yet, which is part of the issue.

Why this is happening now

While this may feel like a bug, it's actually a shift. Big tech companies like Google are rapidly moving toward on-device AI, where your laptop or phone does the processing instead of the cloud.

This shift comes with benefits like better privacy, faster responses and offline functionality. But it also means your devices are starting to store parts of the AI themselves. And that's new (and a storage shock) for many users.

Google responds: what the company actually said

According to Google as reported by Android Authority, the file is tied to on-device AI features powered by Gemini Nano and is downloaded “so features that rely on these on-device models stay ready for use.” The company also said that beginning in February, Chrome started rolling out a way for users to “easily turn off and remove the model directly in Chrome settings.” Once disabled, Google says the model will no longer download or update.

That response addresses one of the biggest complaints users had: the inability to permanently remove the file. Previously, many users reported deleting the folder only to see Chrome automatically reinstall it later.

Still, Google’s statement doesn’t fully answer the larger criticism surrounding transparency and consent. Privacy researcher Alexander Hanff, whose investigation helped bring attention to the issue, argues that users were never clearly informed before the multi-gigabyte download began. Additionally, Chrome’s visible “AI Mode” feature appears to rely on cloud processing rather than the local Gemini Nano model, which raises questions about why the download was happening automatically in the first place.

For now, Google’s position is essentially that the download is intentional, tied to AI features inside Chrome and users now have a manual opt-out option if they don’t want the model stored locally.

Bottom line

This isn’t a “Chrome is secretly installing something malicious” situation, but it does show where software is heading. As AI gets baked into everyday tools, your computer is quietly becoming part of the system that runs it.

The real issue isn’t the 4GB file, it’s whether anyone told you it would be there. Right now, that transparency feels like a work in progress, which makes it worth paying closer attention to what’s being stored on your device.

More from Tom’s Guide

• I realized I was wasting money every time I used ChatGPT — this simple system fixed it
• I used NotebookLM to offload my mental clutter for a week — here’s what I stopped stressing about
• I stopped asking ChatGPT for generic ideas and started using the MrBeast 'obsession' framework instead

Google saves every search you make. Every query you type, every website you click from results, every topic you research — all of it gets stored in your Google account indefinitely unless you delete it.

This search history doesn't just sit unused. Google uses it to personalize ads, refine search results, and build a profile of your interests and behavior. The company can see everything you've searched for years back, creating a detailed record of your curiosity, concerns, shopping habits, and personal questions.

Don't want Google remembering your browsing history? Fortunately, the clearing process takes under five minutes and works from any device with a browser. Here's how to delete your entire Google search history.

More from Tom's Guide

• Tired of Google's AI Overviews? Three easy workarounds to bring back traditional search
• This Google Maps feature saves me 20 minutes every day, and you're probably ignoring it
• Where did your Google Drive storage go? 5 hidden space-hogs you can delete now

Chrome collects data about your browsing activity, and Google isn't entirely transparent about what information it gathers. While the company keeps specific details vague, court filings over the last few years have shed light on some of the data collection happening in the background.

Chrome has faced legal challenges over its tracking practices. A 2024 lawsuit revealed the browser was collecting data even in incognito mode, which led to a settlement requiring Google to delete significant amounts of browsing data.

Chrome's default settings lean toward data collection rather than privacy protection. Many tracking features are enabled automatically, and disabling them requires navigating through settings that most users never explore. Here's what Chrome tracks and which settings you can change to limit data collection.

More from Tom's Guide

• Where did your Google Drive storage go? 5 hidden space-hogs you can delete now
• Bring back Google’s old-school search results with this one simple hack
• Google's new feature lets you customize search to show sites you trust — how to use preferred sources

I spend a good chunk of my day using AI inside a browser — summarizing articles, comparing products, rewriting emails, planning… you name it. And until now, there’s been one constant issue that drives me crazy, maybe you can relate? I keep repeating the same prompts over and over again.

Google just introduced a fix for that. It’s called “Skills” in Chrome, and while the name sounds like a major overhaul, the reality is much simpler — and arguably more useful.

What Chrome 'Skills' actually are

At its core, Skills are reusable prompts built into Chrome’s Gemini sidebar. That means, instead of typing something like: “Summarize this page in bullet points for a busy reader,” you now can save it. When you save a prompt you use frequently as a "Skill," it's saved to run instantly on any page you're viewing.

It’s essentially a way to turn your best prompts into one-click tools that work across the web.

Why this is a bigger deal than it sounds

Before this update, using AI in your browser meant copying text from a webpage, pasting it into Gemini or ChatGPT and re-entering the prompt over and over.

Skills remove that loop entirely. Once saved, a Skill can:

• Work on the page you’re currently viewing
• Apply across multiple open tabs
• Be triggered quickly without rewriting instructions

In other words, you’re no longer chatting with AI — you’re running it with it seamlessly integrated into your workflow.

Even better? There’s also a built-in library of Skills. Google isn’t expecting you to start from scratch. Chrome is rolling out a prebuilt Skills library with ready-to-use options like:

• Summarizing long articles
• Comparing products across tabs
• Extracting key details from a page
• Adapting recipes or instructions

So even if you never write a prompt, you can still use Skills immediately.

How to use Chrome Skills

If you want to try it yourself, simply open Chrome and launch the Gemini sidebar. From there you can enter a prompt you use often or try one of the preset options. If you are using a new prompt of your own, save it as a "Skill" so you can use it again any time. Simply run it on any page with a click

What this is not

There’s been some confusion around this launch, so it’s worth clearing up what Skills aren’t:

• It’s not replacing Chrome extensions
• It’s not a new AI model
• It’s not a major redesign of Chrome

This is a layer on top of Gemini — not a replacement for anything.

The takeaway

The real shift here is how this new feature will change behavior, while completely removing one of the biggest annoyances: repeating yourself.

And if you’re already using prompts daily, that small change can make AI feel a lot faster — and a lot more useful.

Google is moving AI from something you ask…to something that acts. And while this might sound subtle, in practice, it’s a big upgrade, especially if you rely on AI for everyday tasks.

I've already tried a few of the preset prompts and have saved a few of my own. Give it a try and let me know in the comments what you think.

More from Tom's Guide

• I ditched my paper planner for Gemini Live — and it solved the one professional problem I couldn't fix
• I stopped using ChatGPT as a chatbot and turned it into a 'workspace' — these 7 features changed everything
• I replaced my notes app with Gemini’s new Notebooks — and I’m not going back

The MacBook Neo’s 8GB of unified memory has been a major topic of conversation since Apple announced this new budget-friendly $599 laptop. And that makes sense, given how resource-heavy web pages have become, along with Chrome’s infamous high memory usage. If you’re the kind of student or digital nomad this laptop is aimed at, it’s natural to wonder if the Neo can handle your average workflow.

To that end, we decided to put the MacBook Neo through a stress test that simulates a typical day of use. While you might not ever have so many open tabs and apps running on the Neo, it’s interesting to see what Apple’s entry-level MacBook can and can’t handle. At what point will you hit a bottleneck?

Here, we opened Chrome tabs one by one, loading lighter and more demanding websites, and kept going until we started noticing slowdown. To make the test more demanding, we disabled Chrome’s memory saver feature. This is a RAM-saving feature I recommend enabling when using Chrome on the best laptops and best computers, but for this test, we wanted to see how well the Neo would perform without it.

Without spoiling the results, the MacBook Neo is more capable than you would expect from a machine with 8GB of RAM. You need to have a lot going on before you’ll feel the laptop buckling under the workload. Though capable, it’s important to remember that this is a budget machine, so don’t expect to render 4K videos or run other demanding apps without hitting constraints.

Are you worried about the 8GB RAM limit, or do you think it’s enough for the light work the MacBook Neo was designed for? Share your thoughts in the video and in the comments section below. Also, don’t forget to check out our separate gaming test featuring Cyberpunk 2077 and Resident Evil 4 Remake on the Neo.

More from Tom's Guide

• MacBook Neo vs. MacBook Air M2: What’s the best bang for your buck?
• We benchmarked the MacBook Neo vs budget Windows laptops — here's the truth
• We just tested the MacBook Neo vs MacBook Air M5 — which laptop should you buy?

Here's an understatment: The internet is not designed to help you focus. Every platform and notification is engineered for the opposite. I'm guilty of picking up my phone for a 2-step verification, logging on to Gmail and forgetting that I was in the middle of verifying something. Yea...it's bad. Or, it was.

After several weeks of testing dozens of productivity tools, that I didn't even know existed, I've found that AI might finally be the way to get my focus back. I know that sounds crazy considering the amount of AI slop on the internet, but here are the seven AI-powered Chrome extensions that genuinely reduced the bottlenecks and distractions from my work.

TL;DR Quick list

• Best for research: Sider
• Best for note-taking: Glasp
• Best for video/long-form: Merlin AI
• Best for removing visual noise: uBlock Origin
• Best all-in-one assistant: Monica AI
• Best for understanding complex content: Gemini in Chrome
• Best for extracting insights & deep analysis: Claude in Chrome

1. Sider: Best for staying in flow while researching

The “tab-switching spiral” is a silent productivity killer. You open one tab to fact-check and 40 minutes later you’ve got 11 tabs open and no memory of the sentence you started with.

Sider helps break that cycle by bringing answers directly to you. It lives in a collapsible sidebar, responding instantly when you highlight text or need a summary, so you don’t have to leave the page. During testing, I completed an entire research session without opening a single reference tab.

I really like that it collapses and reduces context switching by giving me answers insid my reading flow rather than pulling me away from it.

2. Glasp: Best for turning reading into actionable notes

I'm guilty of collecting bookmarks like I'm some sort of digital hoarder but then never return to them because I forget that I bookmarked them in the first place. Glasp rewires the reading-to-thinking pipeline by turning highlights into an organized, AI-summarized library of ideas. As you highlight articles, PDFs or videos (yes, even YouTube!), it captures key insights and generates summaries you can revisit later.

I noticed my behavior shift from passive scrolling to active extraction — instead of rereading pages, I had a synthesized output waiting for me. This extension replaces passive consumption with active capture to reduce my mental clutter and retain what matters. Any time AI encourages critical thinking, I feel is a win.

3. Merlin AI: Best for quick summaries instead of deep dives

If you're like me, you've clicked a 43-minute YouTube video for a “quick tip” and lose half an hour. Merlin AI summarizes videos, long-form articles and documents in seconds, letting you gauge the value before committing your time. The real benefit in my opinion is it’s intentionality. You read the summary, decide whether the content is worth your attention and move on if it isn’t. What could have become a rabbit hole turns into a 60-second decision.

For me this is a focus win because it takes minutes instead of hours to evaluate the value of content before consuming it. Surprisingly, I find that Merlin has made me more engaged with the content I do decide to consume.

4. HARPA AI: an automation powerhouse hiding in your browser

HARPA AI is a chatbot, but that's only one of its many capabilities. It's more like a behind-the-scenes assistant that quietly handles things like summaries, extracting data and even monitors sites for changes like price drops, restocks or updates, so you don’t have to keep checking back.

The automation is where it really shines. I set it it to handle repetitive tasks, so instead of bouncing between tools and refreshing pages, it does it all for you. If you regularly watch price drops on sites beyond Amazon, this one is worth checking out. It helps to free your attention for other things.

5. Monica AI: Best for keeping tasks moving

I have found that the “context tax” is the hidden cost of leaving your work to ask an AI question. I do it all the time — an idea will come to me out of nowhere and I'll have to leave what I'm doing and open ChatGPT, Claude or something else. But,
Monica AI reduces that tax by living inside your workflow.

Instead of copying text into ChatGPT, opening new tabs or juggling multiple tools, Monica sits in a sidebar and helps you read, understand, summarize and act on information without leaving the page. It can summarize articles, explain complex passages, translate text and answer questions about what you’re viewing, all in context.

Because Monica AI works directly inside the browser, it reduces the constant start-stop cycle of switching apps and re-explaining context — one of the biggest hidden drains on focus for me.

Gemini in Chrome

I use Gemini a lot and have found that Gemini in Chrome helps with everything from search to summaries to clarifying key points. Because the AI help arrives in context, it allows me to stay oriented to any particular task instead of mentally resetting every time I switch tools.

Over the course of a workday, those avoided interruptions add up to noticeably better focus. Less tab-hopping and fewer mental resets is a huge win because I stay anchored to the task while still getting the clarity I really need.

Claude in Chrome

Sometimes I prefer Claude, especially for coding. Gemini and Claude are completely different tools (seriously, they are so much more than chatbots). As with Gemini in Chrome, I can summarize long articles, extract key insights or ask targeted questions about what I'm reading without switching apps or re-explaining context.

When I prefer Claude, it's right there, and that continuity keeps my working memory intact and prevents the subtle bottleneck that builds every time I leave a task mid-flow. Instead of juggling tools, I stay immersed in the material while Claude offers the assistance I need.

Bottom line

I have tried AI browers and can honestly say they aren't for me. I like being apple to uninstall or pick the AI I want in Chrome. That's why after running these produtivity tools for several weeks, I have seen measurable results.

From fewer open tabs to zero 20-minute spirals or 30-second detours, having the AI in my browser assists with productivity in ways I never knew I needed.

AI shouldn't be another thing you have to manage; it should be the reason you have less to manage in the first place. If you’re looking for new ways to stay focused, give these tools a try. You don’t need to install all of them at once. I suggest starting with Sider and Merlin. They address the most common friction points — tab overload and content rabbit holes— with the least amount of setup.

Give them a try and let me know which ones are your favorite.

More from Tom's Guide

• I use the 'reverse brief' prompt to instantly understand any document — here’s how it works
• I use the 'hype check' prompt before every big decision — it's stopped me making 3 huge mistakes
• Gemini 3.1 Pro is a powerhouse for deep work — here are 7 prompts that prove it

Chrome is getting a trio of upgrades designed to help you be more productive: spilt view, PDF annotations and Save to Google Drive. All three appear to be available right now.

Annotations and save to Drive are specifically for PDFs, which we'll cover below. Split view, however, changes how you use tabs and view sites in the browser.

Google announced the new features in a blog post, saying they were developed for both everyday and business users. Here's what they do.

What Split View does

My colleague, Anthony Spadafora, tested out split tabs back when it was still in beta and came away quite impressed with the feature calling it a "browser breakthrough."

"Split Tabs provides a whole new way to browse that, as a long-time Chrome user, has really impressed me so far," he said.

Simply put, split view lets you put two tabs together so you can look at condensed versions of two websites side by side. In your browser, they're also put in the same tab so you can move the double tabs to a new window if you need to.

Now, you can bounce back and forth between the two tabs without having to actually open them or split them between two windows. Within the double tab, you can close one when you're done or separate it once you need more space.

To access it, right-click a tab you're using and click "Add tab to new split view." From there, you can either have it open a blank window that you can use to search or select another open tab to add it to the split view.

PDF Annotations and Drives

The other two upgrades are smaller but PDF-focused and should make editing and saving documents much easier when looking at them in Chrome.

Annotations allows you to highlight text and notes directly in a PDF in the browser. For example, you can use it to e-sign a document without opening a new program or build the feature into a PDF.

To access annotations, you can click the new icon that looks like a squiggly line in the tool bar. Once open you can change the color of your notes and the type of tool you're using.

"Save to Google Drive" lets you save a new PDF directly to your Drive folders without downloading it and reuploading it. When looking at a PDF in Chrome you should see a Drive button with a plus symbol in the tool bar. Clicking that will allow you to automatically save the document.

These saved PDFs will be sent to a dedicated "Saved from Chrome" folder.

All three features are available in Chrome now. If you aren't seeing them, try to updating your browser to the latest version.

More from Tom's Guide

• Google Pixel 10a: Everything we know so far
• AI keeps improving, yet trust keeps falling — here’s what’s driving the backlash
• I use the ‘Onion’ prompt with ChatGPT when I’m buried in tasks — it cuts through clutter in seconds

Google I/O 2026 is still a few months away, but we have an official date from the tech giant and some idea of what we can expect to see.

According to an announcement email sent to Tom's Guide, I/O 2026 will cover AI, Android, Chrome, and cloud "breakthroughs." It also mentions product demos, though the I/O website seems to be teasing more of a software showcase.

Last year, Google held a separate Android Show, but that doesn't appear to be the case for this year's edition.

Google I/O 2026: Date and location

Google has confirmed that I/O conference will kick off on May 19 with the keynote address starting at 10 a.m. PST/ 1 p.m. EST/ 6 p.m. BST. The conference will run through May 20 with Google hosting a variety of talks and demos on several topics, some of which you'll be able to livestream from your home. And if you can't watch the news as it happens, we'll be covering the event in a live blog like we normally do.

As with previous editions, the keynote will be in person and take place at the Shoreline Amphitheatre at Google's headquarters in Mountain View, CA. You should be able to livestream the keynote from the Google I/O webpage, and more than likely on the Google YouTube channel as well.

Google I/O 2026: What to expect for AI

Google hasn't released a schedule for I/O 2026 yet, but there are some things we think the company will discuss including AI.

AI and Google's assistant Gemini have been a key talking point at I/O for the last few years. That is unlikely to change in 2026. Especially as Gemini is being integrated with Apple Intelligence this year, though I doubt Google will address that at I/O.

Gemini will receive much of the focus and we can foresee Google announcing how it will integrate deeper into more programs and devices in the Google ecosystem. Gemini is still mostly absent from cars and wearables, despite it replacing Google Assistant.

Of course, Google has multiple LLMs that are in constant development, so we will hear more about different AI initiatives at Google.

Plus, there's Android which will more than likely feature more AI integration.

Google I/O 2026: What to expect for Android

Speaking of Android. The first Android 17 beta launched last week just in time for Valentine's Day.

The first beta comes with several upgrades like new defaults for running keyboards, navigation and color profiles that should work seamlessly depending on if you're using a foldable or standard smartphone. It also has some new photography modes and better medical device pairing.

We expect Android 17 to launch for Pixel phones anytime between March and the end of June. So it's very likely that Google will have a lot to share about the next Android update at I/O.

We don't know much about 17, but it may take some liquid glass inspiration from Apple.

Outside of phones, we could finally see a real look at the rumored Android/Chrome combined operating system which Google has teased in the last few months.

Last year, Google revealed Android XR, and while I don't expect anything major about the wearable OS, we should hear some updates, especially as new devices are expected to launch later this year.

Google I/O 2026: What to expect for Devices

Google I/O is typically very software focused, but hardware does get some shine. For example, we got an early hands-on with Samsung's Galaxy XR headset at last year's event, then known as Project Moohan.

This year, we might get teases for the upcoming Pixel 11 and the Pixel Watch 5, but I wouldn't put much stock there.

One device we might hear about though is the Xreal Project Aura smart glasses. The partnership was revealed at last year's I/O, and hopefully, we'll learn more at this year's edition.

Google I/O 2026: Everything else we might hear about

Without the schedule, we don't know exactly what to expect but Google did say that Chrome will be featured in its announcement.

We will also likely see more talks and presentations concerning Waymo, cars and Google Pay. Many of those talks may not be livestreamed and will be more technical in nature though.

That said, they could be mentioned in the keynote as something to look forward to.

Tom's Guide will be on hand to track all of Google's announcements and we'll provide more information on the livestreams and features as we get closer to May.

More from Tom's Guide

• AI keeps improving, yet trust keeps falling — here’s what’s driving the backlash
• Get your Pixels ready — Android 17 beta 1 just arrived with a bunch of new features (Update: launch postponed)
• YouTube TV just launched cheaper bundles starting at $54 — here's the four new options

Chrome is getting a major overhaul as Google rolls out a new wave of Gemini-powered features designed to make the browser more proactive.

Announced by Parisa Tabriz, VP of Chrome, the update brings deeper Gemini integration to Chrome on macOS, Windows and Chromebook Plus. Built on Google Gemini 3, the company’s latest AI model, the changes are designed to help users multitask, summarize information and complete multi-step tasks directly inside the browser.

Some features begin rolling out today (January 28), while others are already live or scheduled to arrive later this year.

Gemini gets a permanent place in Chrome

Starting today, Chrome users will see a new Gemini side panel, which keeps the AI assistant accessible no matter which tab is open.

Instead of switching between windows or opening a separate chatbot, users can keep their main webpage in focus while Gemini works alongside it — comparing products, summarizing reviews or pulling together information from multiple sites.

Google says early testers have used the side panel to manage crowded calendars, research purchases and reduce tab overload. The side panel will be available to all Gemini-in-Chrome users.

AI image editing comes directly to the browser

Fans of image generation will appreciate this one. Chrome will now offer built-in access to Nano Banana, Google’s image transformation tool. The feature allows users to modify, edit and generate images on the web using text prompts, without downloading files or opening another app.

Google positions Nano Banana as useful for creative tasks like redesign inspiration or turning research data into visual summaries. The feature rolls out today and will be available to all Gemini-in-Chrome users.

Chrome connects Gemini to Gmail, Calendar and more

Some of Chrome’s Gemini upgrades are already live. Since January 12, Connected Apps have allowed Gemini to pull context from Google services such as Gmail, Calendar, Maps, YouTube, Google Shopping and Google Flights.

That means Gemini can reference old emails, check travel details or help draft messages without leaving Chrome. These integrations are optional and can be managed through Gemini’s settings.

Personal Intelligence is coming later this year

In the coming months, Google plans to bring Personal Intelligence into Chrome — a feature that allows Gemini to remember context from past interactions to deliver more tailored responses over time.

Google says Personal Intelligence will be opt-in, with controls that let users decide which apps are connected and what context is stored. The goal is to make Chrome more proactive, though it also raises familiar questions around personalization and long-term data use.

Auto browse introduces agentic browsing — but only for paid users

The most ambitious addition is Chrome auto browse, an AI agent designed to complete multi-step tasks on a user’s behalf. The feature begins rolling out on January 28 to AI Pro and AI Ultra subscribers in the U.S.

Auto browse can handle tasks like researching travel options, filling out online forms, gathering documents, managing subscriptions and comparing service quotes. It pauses to ask for confirmation before sensitive actions such as purchases or sign-ins.

Google says the feature can also work with images — identifying items in photos and searching for similar products online while staying within a set budget.

To support this kind of agentic browsing, Chrome will also adopt Google’s Universal Commerce Protocol, an open standard developed with companies including Etsy, Shopify, Target and Wayfair to help AI agents interact consistently with shopping sites.

Google says Chrome’s AI is built with new security layers

Alongside the feature rollout, Google published a technical blog detailing how Chrome’s security model is evolving for AI agents.

The company identifies indirect prompt injection — where malicious web content attempts to manipulate an AI agent — as one of the biggest emerging risks. To counter that, Chrome now uses multiple layers of protection, including:

• A separate user alignment critic model that checks whether an agent’s actions match the user’s intent
• New origin restrictions that limit which sites an AI agent can read from or act on
• Mandatory user confirmations for sensitive actions like payments or logins
• Real-time detection of prompt-injection attempts and ongoing red-team testing

Google says these measures are designed to prevent unwanted actions and data leaks while keeping users in control.

Bottom line

Together, these updates signal a clear shift in how Google sees the browser’s role. Chrome is moving beyond displaying information toward actively helping users complete tasks — a direction that mirrors broader industry moves toward agentic AI.

For now, the most visible changes will be the Gemini side panel and image tools, while full agentic browsing remains limited to paid tiers. How useful — or intrusive — these features feel in everyday use may ultimately decide whether this new era of browsing sticks.

More from Tom's Guide

• Anthropic's CEO just warned everyone that the next big AI risk to humanity is 'actually AI companies themselves'
• I didn’t realize Perplexity could do this — 10 hidden features worth trying
• I thought my ChatGPT chats were gone — here’s how I found them instantly

Even with Gmail's category tabs sorting my emails, my inbox still felt chaotic. The real problem wasn't the number of messages but the repetitive process of managing them. I'd open an email, read it, archive or delete it, then click back to my inbox to decide which message to tackle next.

This constant back-and-forth happened dozens of times a day and made checking email feel exhausting. Then I discovered Gmail's Auto-advance feature buried in settings, and it completely changed how I process my inbox. Here's how to enable it and why it makes such a dramatic difference.

More from Tom's Guide

• Gmail lets you unsubscribe from spam emails with a single tap — here's how
• How to stop your personal data from appearing in Google searches
• Tired of Google's AI Overviews? Here's how to bring back traditional search

A new malicious browser extension has been discovered that along with two others, are designed to steal credentials, take over browsers and more. Discovered by security researchers at Koi Security, all three extensions are believed to be created by the same threat actor and put millions of Chrome, Edge and Firefox users at risk.

This most recent malware has been dubbed Zoom Stealer and its uses 18 different extensions to steal online meeting information like URLs, embedded passwords, meeting IDs, schedule times, registration status, topics and descriptions. These extensions affect up to 2.2 million Chrome, Firefox and Edge users and are specifically designed to imitate enterprise level tools for video conferencing apps like Google Meet and Zoom.

The extensions, which succeed specifically because they perform as advertised and have positive reviews, also silently run surveillance in the background to exfiltrate details of the meeting like links, participant lists and credentials using a WebSocket connection. If a users visits a webinar registration page with one of the malicious extensions installed in their browser, the threat actors can harvest all sorts of data including information about the speakers, the hosts and the session metadata.

The Zoom Stealer campaign is the third rolled out from the threat actor known as DarkSpectre. The first, known as ShadyPanda, was found by the same researchers last month and was used to commit fraud, hijack search queries and carry out data theft from more than five million users. The second campaign, known as GhostPoster, used hidden malicious JavaScript code inside logo files in order to gain full control over a browser.

How to stay safe from malicious browser extensions

As with any other malicious extension, if you've downloaded anything that you suspect isn't entirely legitimate, you should remove it immediately.

After removing a problematic extension, you should always reset your account passwords – often, the recommendation is for all your high profile accounts like your email, online banking and social media ones too. Since this could be a hassle, you should use one of the best password managers to make it easier since many can update compromised passwords for you. A password manager can also help keep your passwords organized, safe and can automatically generate strong and unique passwords for each of your online accounts.

I also always recommend using the best antivirus software as well. An antivirus program may not have caught these malicious extensions, but they can scan for malware, spyware and viruses to help you deal with the fallout from a campaign like this one. Antivirus programs also have browser extensions that can warn you when you try to visit suspicious websites, help protect your data with cloud backups and can provide you with a VPN and other extras to add an extra layer of security to protect you when you're online.

Given how effective malicious extensions like the ones described above have been in the past, I highly doubt cybercriminals are going to stop using them soon. Instead, it's up to you to remain vigilant when it comes to downloading and using new browser extensions. When in doubt though, consider whether or not you really need a particular extension before downloading and adding it to your browser.

More from Tom's Guide

• I write about hackers for a living but these are the 3 real-world threats that worry me most — and how to avoid them in 2026
• From misinformation to AI-powered cyberattacks – the top cybersecurity risks for 2026
• I’m a security editor and these are my 3 security New Years resolutions

Google Chrome has dominated the browser market since 2012 when it surpassed Internet Explorer (remember that?) with millions of users across the globe. However, that popularity also means Chrome comes with some limitations when it comes to protecting your privacy, according to a new study

As reported by Neowin, iGaming company Digitain looked at the risks various browsers expose users to in 2025. Chrome isn’t the worst, but its nearly 72% market share means that more people are at risk than those who use smaller and less popular browsers.

You might want to avoid ChatGPT's browser (for now)

Digitain claims that ChatGPT Atlas, the browser from OpenAI, is the worst for your security. Atlas just launched in October, but the team at Digitain felt this browser is apparently designed without security in mind at all.

ChatGPT Atlas failed all state partitioning tests, meaning that the browser does not block websites from tracking what you do across sessions. It ranked low on all three measured metrics (connection and navigation security, privacy and anti-fingerprinting, and tracker and data blocking). Atlas ended with an overall risk score of 99 out of 100.

Chrome was the second-worst browser by comparison, at 76 out of 100. Compared to ChatGPT Atlas though, Google's browser seems like a locked safe

Here’s the list of the 10 worst browsers when it comes to user privacy, along with risk scores (the lower the score, the better the browser):

• ChatGPT Atlas - 99
• Google Chrome -76
• Vivaldi - 75
• Microsoft Edge - 63
• Opera - 58
• Ungoogled - 55
• Mozilla Firefox - 50
• Apple Safari - 49
• DuckDuckGo - 44
• Tor - 40

Microsoft Edge scored better at 63, but Apple’s Safari browser, the second most popular browser in the world, smoked both at 49 out of 100.

“New AI-powered browsers like ChatGPT Atlas and Comet from Perplexity are getting a lot of attention right now, and millions of people are trying them out because of the AI hype,” Paruyr Harutyunyan, head of digital marketing at Digitain, supposedly said in a statement.

“AI works by collecting and learning from data, which means these tools might be gathering more of your personal information than you realize. Just because something uses AI doesn't automatically make it safe or private.”

Digitain noted that Brave and Mullvad Browser have a strong focus on user privacy. Brave has long been built and sold as a privacy-oriented browser, but Mullvad, as the name suggests, was built in partnership between Mullvad VPN and The Tor Project. It’s marketed as an open-source privacy-focused browser by one of the best private VPNs.

Protect your privacy

Chrome likely won't lose the top spot in terms of browser popularity anytime soon, but there are things you can do to ensure your data and privacy are protected. And that carries over no matter which browser you use to surf the web.

First, you’ll want to deny or restrict permissions for things that aren’t necessary like location, contacts, and phone access. Likewise, you can turn off personalization settings in the browser's desktop or mobile app. For example, in Chrome, you can turn off the browser's ability to sync with your Google account as well as disable add personalization and activity tracking.

For things you do allow, like storage access (to save downloads) and camera and microphone access for video calls, make sure to recheck your permissions monthly, especially if any new ones have been added.

To protect your privacy even further, you might also want to consider using one of the best VPNs when browsing the web. Doing so will keep your data encrypted while masking your IP address. Good VPNs won't log or share your data and verified no-logs policies can back up these claims. Some VPNs even offer tracker, ad and malware protection.

Additionally, you should consider using one of the best ad blockers, too. Google has cracked down on ad blockers recently, but there are still some options that work available. Using one means you block some tracker activity and reduce the number of annoying ads you see.

Finally, if you're are a Chrome user, you might want to consider switching browsers and start using something like Mullvad or Brave as, at least according to Digitain, they're safer for your privacy compared to the Google elephant in the room.

More from Tom's Guide

• LG just announced its first Micro RGB TV — and it's a massive upgrade for Mini-LED
• Best LG TVs: the best LG OLEDs tested and reviewed
• No more cables! Dolby and LG are making wireless Dolby Atmos sound happen in 2026

Google’s Tuesday update for Chrome includes 13 security fixes, including patches for four high severity bugs. Given that the world's most popular browser has 3.4 billion users, these bugs could potentially affect billions of people if left unpatched.

Though details about the flaws are, as usual, relatively scarce as Google waits for users to update their browsers, we do know that one of the flaws was discovered within Chrome's Digital Credentials feature. This is a tool that allows users to share verified information from a digital wallet with websites in order to prove their identity across devices.

Another one of these vulnerabilities (tracked as CVE-2025-13633) could allow a remote attacker to potentially exploit a heap corruption flaw via a maliciously crafted HTML page. What this means is that an attacker who lured a potential victim to a fake site containing specially crafted code with a bad extension, could steal information handled through digital credentials. Attackers could then access information which would normally require a passkey.

How to stay safe from browser security flaws

The scope of these flaws and the browser popularity – 3.4 billion users – is precisely why its so important for everyone to update their software (operating systems, security software, apps and browsers) promptly. If you don’t patch holes in your system, it’s easier for attackers to exploit them when you go online, even if you're just browsing the web.

Also, make sure you don’t delay restarting your browser. For these fixes, the latest version number is 143.0.7499.40/41 for Windows and macOS and 143.0.7499.40 for Linux. If you let Chrome update automatically, you're already protected. However, if you never close your browser or turn off your computer, then you might be running an older version of Chrome which puts you at risk.

To update manually, click the More dots (three dots in the upper right corner), then go to Settings > About Chrome. If there’s an update available, Chrome will download it. Restarting Chrome will install the update and you’re all set.

Some of these fixes also apply to other Chromium-based browsers as well, so keep an eye out for updates for Brave, Edge or Opera. And while you're at it, make sure you run a scan using the best antivirus software to ensure your computer is virus free.

More from Tom's Guide

• New Android banking trojan lets hackers stream a live feed from your phone and control it in real time — how to stay safe
• Google just fixed 107 security flaws including two zero-days — update your Android phone right now
• Over 4 million users hit with spyware that can turn your browser extensions into malware — how to stay safe

With AI browsers suddenly everywhere, I have held my stance that Chrome with Gemini is all anyone needs. But I decided to do a hands-on test to see if my opinion held up with nine real world searches.

ChatGPT Atlas, the new web browser from OpenAI, is built directly around the ChatGPT assistant and is designed to combine web browsing with AI‑driven tasks like page summarization, research and actionable automation. It's currently only available on macOS.

Gemini in Chrome embeds the AI‑power of Gemini 3 directly into the Google Chrome browser, meaning you can ask questions, summarize articles or compare open tabs without ever leaving your browsing session.

So I stress-tested both tools the way real people actually use them: hunting for Black Friday deals, planning travel, finding events, researching news, cooking dinner and more. What I discovered is that these two assistants have fundamentally different personalities. Atlas tries to think with you: it remembers context, shapes decisions and guides you toward an outcome. Chrome, on the other hand, works more like a turbocharged search engine — incredibly fast, detailed and structured, but sometimes stopping short of the finish line.

After nine hands-on tests, clear strengths emerged on both sides — and some surprising weaknesses.

1. Shopping Black Friday deals

Prompt: Find me a Black Friday deal on a 55-inch smart TV under $400

ChatGPT Atlas curated a list of deals and included a “Top 3” list with clear pros and cons for each, plus, actionable guidance to support different needs and budgets.

Chrome with Gemini 3.0 presented a comprehensive but somewhat overwhelming list of options without a clear prioritization and no guidance whatsoever. It gave me info like a product search engine rather than an assistant.

Winner: ChatGPT wins because it took a product list and turned it into a concise, easy-to-use buying guide, directly helping me make a decision.

2. Summarizing news

Prompt: Summarize the latest news about OpenAI in 3 bullet points

ChatGPT Atlas offered a longer list of recent developments, but the information provided was less cohesive, mixed major strategic moves with minor trademark issues and exceeded my requested three bullet points.

Chrome with Gemini 3.0 delivered a concise, three-bullet summary that successfully grouped related developments under clear, high-level themes, providing a more strategic and digestible overview of OpenAI's current direction.

Winner: Chrome wins for adhering precisely to my request for three bullet points and synthesizing the information into a clearer, more strategic summary.

3. Recipe help

Prompt: Find a 30-minute dinner recipe I can make with broccoli, rice, and eggs

ChatGPT Atlas was more thorough by including helpful tips (like using day-old rice), offering ingredient alternatives and breaking the instructions into clear, distinct steps.

Chrome with Gemini 3.0 gave me a functional recipe with clear, concise instructions, making it a solid, no-fuss option.

Winner: Chrome wins because the recipe was much easier to follow and did not require extra steps. The images of the recipe on the side were a helpful addition.

4. Local things to do

Prompt: What’s a fun holiday event happening in NYC this weekend?

ChatGPT Atlas focused on a single, classic event with helpful logistical tips, but this narrow scope is less helpful for someone looking for options. This makes me wonder if ChatGPT was going off of memory, knowing that I have young children that may not enjoy sitting through holidays shows in a theater.

Chrome with Gemini 3.0 delivered a wide-ranging list of diverse events across multiple categories (theater, markets, family activities), providing specific venues, times, and dates, which is far more practical for planning a weekend.

Winner: Chrome wins for offering a variety of concrete options, allowing me to choose rather than just assuming what I would want. The options allowed for me to pick based on who I was with and our interests, rather than just highlighting one.

5. Research and information

Prompt: Open the official website of MoMA in New York, identify the current exhibitions listed on the homepage, and summarize the three most prominent ones in one short paragraph each. Include: exhibition title, artist(s), theme, and closing date.

ChatGPT Atlas provided three distinct, well-crafted paragraphs that fully include the requested elements — exhibition title, artist(s), theme, and closing date — with richer descriptions of each exhibition's concept and significance.

Chrome with Gemini 3.0 correctly identified the three exhibitions with their titles and closing dates in a single, concise paragraph, but it lacked the thematic depth and descriptive detail requested.

Winner: ChatGPT wins for delivering more comprehensive and insightful summaries that better fulfill the user's prompt.

6. Trip planning

Prompt: Compare flights from NYC to Miami next month — what’s the best deal?

ChatGPT Atlas delivered a superior, multi-faceted comparison by breaking down options into clear categories (cheapest vs. best value), explicitly comparing airports (MIA vs. FLL) and providing crucial context about trade-offs like baggage fees and comfort.

Chrome with Gemini 3.0 correctly identified the cheapest airlines and general price range, offering a solid starting point for a budget-conscious traveler like me.

Winner: ChatGPT wins because it offered a more nuanced and practical guide that helped me make an informed decision based on my specific priorities, not just the lowest sticker price.

7. History lesson

Prompt: Tell me the origin of Thanksgiving in the U.S.

ChatGPT Atlas offered a well-written, engaging narrative that covered the key historical points, including modern context, in a flowing, conversational style.

Chrome with Gemini 3.0 delivered the information in a more organized, textbook-like structure with clear thematic sections (Early Observances, Becoming a National Holiday, Modern Holiday), making the historical progression and key figures easier to follow and reference.

Winner: ChatGPT wins because it presented the information in a more engaging way without feeling like a history lesson with a structured format.

8. Creating an Amazon wishlist

Prompt: Help me build a holiday wishlist on Amazon — add 5 gift ideas under $25

ChatGPT Atlas immediately fulfilled my request by providing five specific, on-budget gift ideas with brief, appealing descriptions. With memory enabled, it provided a list that makes sense for me.

Chrome with Gemini 3.0 correctly explained the mechanics of how to create an Amazon wishlist but failed to provide the requested gift ideas, making the response incomplete and unhelpful. This was surprising considering I have personalization enabled.

Winner: ChatGPT wins because it directly answered my core request for gift ideas, whereas Chrome only addressed the procedural aspect of the task.

9. Book suggestions

Prompt: Find 3 well-reviewed books by female sci-fi authors and show Goodreads ratings

ChatGPT Atlas offered a solid list of famous, highly-rated books, but included The Handmaid's Tale, which is often debated as dystopian fiction rather than core sci-fi, and The Power, which has a noticeably lower rating.

Chrome with Gemini 3.0 curated a list of three undisputed sci-fi classics that are not only all highly-rated (all above 4.1) but also represent a diverse range of subgenres and thematic depth, from sociological exploration to space opera.

Winner: Chrome wins for selecting a more cohesive and consistently top-tier list of science fiction novels, with stronger and more uniform ratings.

Overall winner: ChatGPT Atlas

After spending several weeks using both tools side-by-side, my view has changed more than I expected. Chrome with Gemini is still powerful, fast and incredibly convenient — especially because it lives inside the browsing environment most of us already rely on.

But in real-world, task-driven use, ChatGPT Atlas proved itself more frequently capable of guiding me to an actual decision, not just giving me information. Atlas remembers context, adapts to my needs and shapes the results into something usable far more often than Chrome does.

That said, Chrome is still more widely available, while ChatGPT Atlas requires me to switch computers and I can't use it on my phone yet. Plus, Chrome excelled at structure, speed and breadth, particularly when I needed a high-level view of news or options. In a few cases, it even outperformed Atlas by being more focused and concise.

But when I look at the nine tests as a whole, a pattern emerges: Chrome is a brilliant information engine; Atlas is a proactive assistant. And depending on what you need, either could be the better pick.

More from Tom's Guide

• I just tested ChatGPT-5.1 vs. Grok 4.1 with 9 prompts — and there's a clear winner
• ChatGPT-4o vs. ChatGPT-5.1 — I tested both and the winner surprised me
• I tested ChatGPT vs Gemini vs Claude to see which chatbot is the biggest people-pleaser — one went way too far and compared me to Steve Jobs

AI browsers are everywhere right now. Perplexity has Comet. OpenAI has Atlas. Even smaller players are trying to reinvent how we find and use information online. I’ve tested all of them this year, and they each have clever uses, especially for people who want conversational answers instead of sorting through dozens of tabs.

ChatGPT Atlas is designed to replace traditional search with fast, conversational answers and deep, synthesized reasoning. It can analyze webpages, summarize long documents and pull in real-time information directly inside the browser interface, acting like an AI research assistant.

Similarly, Perplexity’s Comet delivers conversational search results with cited sources, live web data and clean, structured explanations.

But after using Gemini 3 inside Chrome for the past week, I realized something: I don’t actually need a dedicated AI browser anymore. Google has effectively turned Chrome into one, simply by upgrading Search.

Gemini 3 now powers a new AI Mode in Google Search — and because most of us (3.83 billion internet users) use Chrome by default, those capabilities show up right where we already spend our time. No switching to a different browser. No new interface to learn. No rebuilding an entire digital workflow from scratch.

Here’s what Gemini 3 can actually do inside Chrome today and why it covers nearly every major feature that AI browsers are trying to sell.

1. Gemini 3 gives Search deeper reasoning — right inside Chrome

AI Mode in Google Search now uses Gemini 3’s advanced reasoning abilities.
When you ask a complex question, Search automatically breaks it down into smaller sub-queries, finds relevant sources and synthesizes everything into a clear answer.

For Chrome users this means more contextually aware responses and better structureD explanations. It also means fewer irrelevant links dominating the page and the ability to ask follow-up questions without restarting your search. This is the core feature AI browsers advertise, yet now it’s built directly into the search engine Chrome already uses.

2. Chrome now supports multimodal search powered by Gemini 3

Since it's launch, Gemini 3 brings a major upgrade to how Search understands information. It can now interpret images you upload including screenshots, photos of objects and text combined with visual content.

While Perplexity and Atlas support some multimodal features, Google Search has a huge advantage: it’s built on top of Lens, global search data, and Google’s existing systems.

Multimodal search works on mobile and desktop through Search; there's no need for a separate AI browser window to do the same thing.

3. Real-time camera search (mobile) lets Gemini 3 analyze things you point at

This is one of the most impressive new features. On mobile devices, you can activate AI Mode in Search and share your camera feed in real time. Gemini 3 can analyze what it sees, answer questions, identify objects, explain steps or troubleshoot issues.

This is Search with live vision, and AI browsers don’t currently offer anything comparable.

While this feature is mobile-only for now, it still lives inside Search, meaning you access it through Chrome without switching tools.

4. Webpage summaries and insights are now native in Search

AI browsers often highlight webpage summaries as their standout feature, but Gemini 3 brings the same capability directly into Search.

You can ask:

“Summarize this page”

“Give me the key points”

“Explain this step by step”

Search generates a condensed, structured overview based on the actual page content, using the same AI that powers AI Overviews. And because it’s Google, you also get clear citations and links to the sources behind the summary.

Again, no need for a new browser. It’s all built into Chrome’s default search experience.

5. Planning tasks and multi-step reasoning work naturally in Chrome

Gemini 3 isn’t just summarizing pages. AI Mode can handle more structured tasks, like:

• Planning a trip
• Creating a recipe timeline
• Comparing products
• Drafting research outlines
• Breaking down tutorials

This is one of the big selling points of Perplexity Comet, yet Gemini 3 offers the same workflow directly in Search, which you already use every day. Instead of switching to another AI tool, just use what you have because the Gemini 3 upgrades have made it smarter.

6. Chrome keeps everything an AI browser can’t replace

Even the most polished AI browser struggles with one big problem: you have to abandon the browser you already live in. What I mean is, all your bookmarks are probably already in Chrome. When you download a new AI browser, you have to switch those over. And it's not just bookmarks, you'll also have to switch your extensions, your autofill and passwords, your synced history, your tab workflow, your profiles and your workspace integrations.

Gemini 3 upgrades your existing setup instead of asking you to start over in a new environment. That’s a massive advantage for me. And one that’s easy to underestimate until you try moving everything to another browser.

7. Gemini 3 appears everywhere you already use Google — not in a silo

Anothre reason, at least for me, AI browsers suddenly feel less essential. Gemini 3 shows up across:

• Search in Chrome
• the Gemini app
• Android system features
• Google Workspace apps (with more coming)

It’s ambient across the Google ecosystem, instead of locked inside a standalone browser window. AI browsers require you to go to them. Gemini 3 meets you where you already are.

Bottom line

What's that saying, "Don't fix it if it isn't broken?" That's the way I feel about Google. Sure, AI browsers like Perplexity Comet and OpenAI Atlas are pushing the industry forward, and they deserve credit; they helped make conversational search mainstream. But with Gemini 3 powering AI Mode in Search, Chrome has effectively become an AI browser without forcing anyone to download something new.

If you already live in Chrome, Gemini 3 gives you the same conversational answers, deep synthesis, multimodal reasoning and planning tools, all in the browser you’re using anyway.

For me, that’s the deciding factor: I don’t need a new browser because Chrome continues to have everything I need, plus it's now smarter with AI.

• ChatGPT-4o vs. ChatGPT-5.1 — I tested both and the winner surprised me
• I just tested ChatGPT Atlas vs. Chrome with Gemini 3.0 — and there's a clear winner
• I just tested Gemini 3 vs ChatGPT-5.1 — and one AI crushed the competition

More from Tom's Guide

In a security advisory published on Monday, Google released details about an emergency security update that was issued to fix the seventh zero-day vulnerability exploited in attacks against Chrome this year. The zero-day in question (tracked as CVE-2025-13223) is categorized as a high-severity vulnerability and there's also an exploit that has been used in the wild, according to the company.

Google has fixed the flaw, and new versions will roll out to users via the Stable Desktop channel throughout the coming weeks, however a patch may be immediately available if you check for updates. Though the browser does automatically update whenever security patches are made available, users can make sure they’re installed by going to Chrome > Help > About Google Chrome and then clicking Relaunch once the installation process is complete.

As usual, though Google has confirmed that this flaw has been used in the wild to carry out attacks. However, it has not issued any further details, stating: “Access to bug details and links may be kept restricted until a majority of users are updated with a fix. We will also retain restrictions if the bug exists in a third party library that other projects similarly depend on, but haven’t yet fixed.”

This is typical, as sharing details is something the Google avoids as it may encourage further exploits of the zero-day before enough users apply its patch to fix it. The bug, which was reported by a member of Google’s Threat Analysis Group (TAG), is caused by a type confusion weakness within Chrome’s V8 JavaScript engine. If exploited, it could allow an attacker to achieve arbitrary code execution or program crashes via a maliciously crafted HTML page. It makes the seventh zero-day exploit patched this year, with the other fixes being patched in March, May, June, July and September.

How to keep your browser secure

This is exactly why it's so important to make sure that your software and operating system are kept up-to-date. You can always set them to update automatically in order to make things easier for you.

Likewise, the best antivirus software will often also let you automatically schedule scans to help protect your system from malware and other viruses. When you install one, also make sure to set up all of the extra features like a VPN, a hardened browser or other protections that can keep you safe while browsing.

Given this year's track record, it's unlikely that this will be the last zero-day exploit we see in 2025, so you may as well ensure that your browser is set to update automatically.

More from Tom's Guide

• DoorDash was just hit with its third data breach – what to do next
• Black Friday shoppers under attack from AI-powered scams — here's how to spot them before it's too late
• Don't risk it — get your annual antivirus coverage for less than $25 with these Black Friday deals

Your digital footprint is likely much larger than you realize. Every social media post, account signup, and online purchase adds to the trail of personal information scattered across the internet. While completely disappearing from the digital world is nearly impossible, you can significantly reduce your online presence with a few targeted strategies.

Whether you're concerned about privacy, trying to minimize targeted advertising, or simply want a fresh start, these methods will help you take back control of your personal information.

While these steps won't completely erase your digital existence, they will significantly reduce your online footprint and make your personal information much harder to find. The key is persistence. Online privacy isn't a one-time effort but an ongoing process of monitoring and maintaining control over your digital presence.

More from Tom's Guide

• How to stop your personal data from appearing in Google searches
• Do you need to use a data removal service?
• I’m a security editor and this is one security rule I never break

Is this a safe space? Can I be honest? The internet is pure chaos right now — a constant cycle of doomscrolling, sites packed with AI slop that are built to be as addictive as possible, dark patterns (that cost Amazon $1.5 billion) and a news cycle that can reap pure havoc.

In February, the answer to my prayers came in Opera Air: the world’s first mindful browser. And in 2025, a browser that can take you out of this noise, strip out all the UI clutter of Google Chrome and provide mindfulness is needed now more than ever.

Even better, over the past 12 months, Opera has added more features to really add an air of calm to your browser, as well as a downright adorable cat companion.

Does it sound kind of gimmicky? I feared so at first, but over the course of a year, I’ve fallen in love with Air to the point that it pried my hands off of Google Chrome. Not even Opera’s breakthrough AI agentic browser Neon could tear me away from the zen vibes of using this on my day-to-day.

It’s a browser that wants you to…stop browsing!?

The first thing that immediately jumped out at me (aside from the gorgeous new dark “Obsidian” theme being far easier on the eyes), is the fact that Opera Air seems to be hellbent on making sure I don’t browse too much. Like, isn’t this supposed to be the business model of a web browser — clocking hours of website engagement?

That’s where Take a Break comes in, and in a sea of browsers begging for your time, it’s a serious reprieve. Here, you’ll find a range of guided mindful exercises you can take that come with a nice spa-like gentle sound track:

• Breathing: A simple breathing exercise to help reduce stress
• Neck exercise: Guided neck rotations that really helped ease the ache I feel at the back of my neck
• Meditation: This came in clutch when switching to a big, daunting project task — focusing myself and getting into a more relaxed state
• Full body scan: This one took me a while to figure out and notice the sensations in my fingers and toes, but once I did, this formed a part of my end-of-day routine

On top of that, you may see the three line icon gradually empty over time. This is the body battery indicator, and it is a visual indicator of when you’re probably feeling drained and you can get prompted to re-center yourself with an exercise. Think of it like those Mindfulness prompts on your Apple Watch, only a whole lot less annoying!

Binaural focus

In my Chrome era, Lo-Fi girl was always open in a tab with chill vibes to help me focus. But once I donned a pair of good headphones and got into the world of binaural Boosts, there was no turning back.

For those uninitiated, binaural tones are bass notes that are played at different frequencies to encourage certain subconscious reactions. For example, when I’m hard at work typing a story (like this one right now), I fire up the ‘Energized Focus’ track with a 12Hz tone to really help me lock in. As I come to the end of the day and I need to sign off both physically and mentally, a 30-minute session with ‘Deep Relaxation’ 5Hz tones is just what the doctor ordered.

Over the summer, another update was added that provided a track EQ of sorts — meaning you could turn off the background music and sound effects to get nothing but the bassy binaural tone. Pair that with your own tunes on Spotify or Apple Music and you can come up with some amazing combinations.

My favorite personally? Well, throw on ‘Deep Relaxation’ and listen to Loyle Carner’s ‘Not Waving, But Drowning’ and thank me later!

The Hygge browser

Pronounced “hooga” (yes, I was just as confused as you reading that word) is the Nordic concept of being content simplicity — focusing on the smaller details and removing unnecessary distractions to be present in the moment.

That is exactly what Opera Air has been built around, and it shows in its super clean UI and thoughtfully-placed exercises in mindfulness and focus. Not even the animated cat could get me down (mostly because I’m a dog person).

Once I saw this “mindful browsing companion,” I was skeptical. I had Catz on my Windows 98 PC, and even at 8 years old, I found it annoying. But it’s not intrusive on the UI, and can actually be a cute distraction when I needed it!

Plus, you’re getting all the nice-to-haves that Opera is well known for in its browsers — built-in VPN and ad blocker, multi-modal AI that can understand page context and even support offline LLM usage, and all your message, music and social services in a bar to the left.

The internet is pure chaos right now, and Opera Air is your best line of defense against it.

More from Tom's Guide

• I just tested the new ChatGPT Atlas browser — here’s 4 big ways it beats Chrome
• Opera Neon wants $20 a month to replace Google Chrome — is it worth it?
• Hackers can use prompt injection attacks to hijack your AI chats — here's how to avoid this serious security flaw

OpenAI’s new browser, ChatGPT Atlas, brings AI directly into your daily web routine. But for someone like me — who rarely uses their desktop to browse and doesn’t need constant handholding to read the internet — Atlas feels like a co-pilot I didn’t really ask for.

Don’t get me wrong, Atlas is seriously impressive. It's essentially a browser with ChatGPT baked in; it even looks and acts exactly like ChatGPT. With quick page summaries, personalized context and memory-aware suggestions, it’s easily the most intelligent browser I’ve used.

But here’s the thing, after trying it with several prompts, I realized I don't really need it. I'm a capable browser. I know how to search, skim, click and backtrack without AI. For the record, I’ve also tried Comet and didn’t care for it either. Chrome is the only browser that feels useful. Here’s why.

I usually browse from my phone, not my desktop

My desktop is for working. Just as I don’t sit at my desk and scroll Instagram on my phone, I don’t sit on my desktop and search the web. If I ever need to fact check something, I usually go to the link directly. I certainly don’t need ChatGPT Atlas to show me how to get there.

In other words, the setup alone makes Atlas a bit of a mismatch for me. The entire experience is desktop-only for now, and while it works well in that space, it’s not where most of my browsing happens. On mobile, I’m tapping through links quickly, while sitting in the school pickup line or waiting for my turn at Starbucks. For that reason, until Atlas makes it to mobile, I won’t be using it.

ChatGPT is harder to interrupt — but thorough when the answers matter

If I’m on Chrome or using Google, I can shift gears before the links populate. With ChatGPT Atlas, it will sometimes go on (overboard!) with a response. The follow up question is also a frustration for me. I could hit the black square to make it stop mid-query, but even that isn’t as easy as just typing in something new with Google.

One thing I will say is that I do find it helpful to have ChatGPT available for follow-up questions. For example, I looked for lights for reading in bed and ChatGPT was right there to help me find the best deal. That was much easier to do than if I had been Googling it, making the response feel much more personalized.

I also appreciate seeing citations within the chat, unlike Google’s AI Overviews, which often feel like the AI is saying “just take my word for it.”

Still, I don’t need this much help while surfing the web

I hate to admit it, but I’m older than the internet. I can still remember the first thing I ever Googled. I don’t need a tour guide every time I open a new page. I don’t need definitions of terms I already understand or summaries of articles I’m already reading. I enjoy reading, it’s literally why I click on the articles, which, by the way, I prefer to do from my phone.

For some people, Atlas is a dream — students, researchers or anyone dealing with information overload. But if you’re a confident browser with decent reading comprehension and a clear sense of what you’re looking for, it might feel like overkill.

What Atlas does well (that Chrome doesn’t)

Here are the genuinely useful tools Atlas brings to the table:

• Thread memory lets you keep multiple searches or rabbit holes organized across tabs.
• Context-aware chat knows what you’re reading and helps you dig deeper.
• In-browser writing tools make drafting emails, summarizing docs and brainstorming seamless.

If your workflow is already integrated with AI, these are game-changing. As much as I use AI, these features just aren't necessary within my browser. I like having them separate.

Bottom line

More from Tom's Guide

• Claude Memory just launched and it syncs with ChatGPT — here’s how it works
• I wore Friend, the AI companion that listens all day — here’s what surprised me most
• I tested Windows 11’s new AI upgrades — and these 3 are actually worth using (plus, one I can't wait to try)

The wait is finally over: OpenAI has released its own browser, ChatGPT Atlas, in a daring bid to finally dethrone Google Chrome.

The speculation has lasted for the better part of a year, and now, we can all finally test whether ChatGPT Atlas is the long-awaited Chrome replacement many hoped for—provided you own one of the best MacBooks or a computer running macOS. After watching OpenAI’s latest livestream announcing its new browser, I immediately turned on my Mac mini M4 to find out.

While many jump between browsers at a moment’s notice (such as Jason England jumping on Opera Neon), I’ve been a Chrome diehard since Google’s browser launched in 2008. At the time, switching from Internet Explorer on PC was a no-brainer, but Chrome hooked me so much that I even installed it on my old Mac to replace Safari.

Here are my initial thoughts on ChatGPT Atlas after using it for a full day including what I love and what I hate about OpenAI’s new browser.

What it does better than Chrome

Quick sidebar access to ChatGPT

Although Google recently made Gemini directly accessible in Chrome, I honestly prefer OpenAI’s approach with ChatGPT in its Atlas browser. Gemini’s way of working — which often means a floating window popping up and distracting you — feels overpowering.

With ChatGPT, the experience is more discreet. Instead of a big pop-up that can cover the content you're reading or looking at, the chat is right there as a consistent sidebar. To start chatting with ChatGPT on any page, you just click the little button in the top right corner of Atlas.

For those of us who rely on our favorite AI chatbot throughout the day while working, studying, or researching, a small design difference like this can really matter. This Atlas approach means ChatGPT is there exactly when you need it without taking you out of your current flow, and frankly, I could really get used to this kind of seamless, non-disruptive integration.

Full-sized and scrollable tabs

If you’re like me, you know how quickly research spirals out of control. I can easily end up with a dozen or more tabs open across two separate, side-by-side browser windows, and that’s where Chrome’s tab problem begins. Your tabs get smaller and smaller until eventually, all you can see is a tab with an "X" in it. Since Chrome is my most-used app, I just sort of lived with that reality. That is, until I tried ChatGPT Atlas.

While Atlas’s default tab layout looks like Chrome’s, the real game-changer is the option to switch to Scrolling Tabs. This single setting made a massive difference during my short time with the browser. Instead of those teeny-tiny tabs, you get full-sized tabs that never shrink. When you open more tabs than can fit, you just hover your mouse over the tab bar and scroll to reveal the rest.

This works seamlessly whether you're using your mouse’s scrollwheel or your laptop’s trackpad, making navigation lightning-fast. Scrolling Tabs are simply awesome. Honestly, if Google were to add this feature to Chrome — even if it’s just tucked away behind an experimental flag — I wouldn't argue in the slightest.

Full URLs are just a toggle away

I know plenty of people prefer not seeing a page’s full URL, and I get it — it gives your browser a cleaner look. But I’m not one of them. Often, I really need to see the full address. Whether it’s when I’m writing deals and need to cut off tracking nonsense at the end of an Amazon link, or if I want to inspect a URL for a cybersecurity story I’m working on, seeing the whole thing is critical for my workflow.

That's why I appreciate that Atlas puts this option front and center. Instead of hiding this setting behind multiple clicks or deep-dive flags like Chrome does, it’s one of the first options you see in the browser’s Settings menu under the General tab. The best part is that OpenAI uses a simple toggle switch for this, making it a breeze to enable when I need to look at a URL closely and just as fast to turn off when I don't.

It’s Chromium-based because of course it is

To Microsoft’s credit, when it finally retired Internet Explorer for Microsoft Edge, the software giant repeatedly said that its new browser was Chromium-based. Maybe this was to convince more people to switch (it certainly worked) or perhaps they were just giving credit where it was due.

However, Sam Altman and the crew at OpenAI's ChatGPT Atlas announcement didn't exactly lead with that detail or even mention it at all. The minute they opened an incognito window on stage, though, myself and countless other tech nerds instantly noticed that familiar look.

But why should you care that Atlas runs on Chromium? It actually comes with a couple of huge perks. For starters, if you have a favorite browser extension from the Chrome Web Store, you can instantly add it to Atlas without any fuss. It just works. Likewise, given how much work Google and the Chromium Project put into constantly updating the code, you can rest easy knowing you won't be dealing with nasty vulnerabilities. That enormous development effort means Atlas gets the same security patches as Chrome, and will likely get them fast.

How it falls short compared to Chrome

No seamless switching between Google Accounts

If you have a personal Google Account and also use one for work, then you know how simple it is to switch between the two in Chrome. Yes, you can click on your profile icon in the top right to do so, but you can also use one of my favorite keyboard shortcuts (Ctrl + Shift + M) to do the same thing instantly.

After installing Atlas, my personal Google Account was available after signing in, but that’s where the convenience ended. I looked everywhere — and I mean everywhere — but there was simply no option to add a second Google Account.

For my day-to-day workflow, this is a deal breaker and why I won’t be ditching Chrome anytime soon. You can’t properly manage your work and personal digital lives without this basic feature which is exactly what this fledgling browser needs to be a serious contender. However, OpenAI could easily add this feature in an update and I’ll keep my fingers crossed it does.

Constant reminders to upgrade

One of the best things about Chrome is that, like other Google services, it’s completely free. Sure, you can pay for extra storage in Google Drive, but the reason products like Chrome and Google Docs are so ubiquitous is because they’re free and easily accessible. Google makes its money elsewhere.

Don’t get me wrong, OpenAI’s ChatGPT Atlas browser is also free. However, the pressure to upgrade felt constant while using it. For instance, after I finally, and begrudgingly, made it my default browser, the "Make Default" prompt at the bottom of the home page didn't disappear — it instantly flipped to a jarring "Upgrade" button. That was definitely not what I was expecting.

I get it: like any other AI company, OpenAI needs to make money. But these constant, subtle nags to upgrade I saw throughout Atlas didn't convince me to open my wallet. Instead, they were distracting and honestly, a real turn off that hurts the experience of this otherwise promising new browser.

More from Tom's Guide

• ChatGPT Atlas is here — how to download for macOS now
• 7 prompts I use for every AI chatbot — and they work for just about everything
• ChatGPT Projects completely changed the way I organized my ideas — here's how

Gemini can now summarize any webpage directly in Chrome without needing to open the Gemini app. The feature works on Android devices, making it faster to get quick summaries of articles, documentation, or any content you're viewing in your browser.

Previously, summarizing a webpage required copying the URL, opening the Gemini app, and pasting the link into a query. Now it's a single button press that generates an instant summary of visible content on the current page.

Here's how to use Gemini's page summarization feature in Chrome and what to expect from the summaries it generates.

More from Tom's Guide

• I finally figured out how to turn off Google's AI Overviews
• 3 tips for creating amazing images in Google's Nano Banana
• How to use Google Gemini's Guided Learning feature

When it comes to notifications, it's overwhelmingly easy for your lock screen to get swarmed with little boxes pinging for your attention. It can easily drown out information you actually want to see.

Google seems to be paying attention and has released a new tool for Chrome on your phone meant to help quiet the number of notifications you receive. And it's based on a feature that already exists.

Android 11, released in 2020, introduced a feature that revokes notification permissions from apps you haven't opened in a while. Essentially, Google is adding this feature to Chrome Safety Check, launched in 2024, which was built to "give you even more control over your personal data."

At the time, it would remove camera, microphone, and location permissions for sites and auto-detect spam, but it didn't do so for notifications.

Similar to the deep sleep ones that Android occasionally sends you. The Chrome version is an expansion of that for sites you don't visit often. It will automatically revoke permissions for notifications.

According to Google, Chrome on Android will now stop notifications from websites that you haven't interacted with in some time. It's not clear how long that window stays open until Safety Check snaps it shut.

This only applies to normal websites, not installed apps or web apps. Google says that Chrome will notify you when it's doing so. When it comes to apps, you can already choose to disable most notifications in your phone's settings menu.

Notification overload hurts engagement

Google says the reason it introduced the feature is that constant notifications actually hurt engagement.

"Data indicates that users frequently receive a high volume of notifications, resulting in minimal engagement and high disruption. Less than 1% of all notifications receive any interaction from users," the company's blog reads.

Of course, if Chrome gets it wrong, you can always re-enable notifications or turn the tool off. The feature should work on desktop as well if you receive notifications on your computer. For a while now, I've made it a personal habit to block all sites from notifying me of anything as soon as I'm prompted. The banned list is long.

I have yet to receive any silencing notifications from Chrome on Android, but Google tends to roll out launches, meaning that not everyone may have received the feature just yet.

In the meantime, here's how to check out your Safety Check: Open Chrome > click the three-button action menu > navigate to Settings. From there, scroll down to 'Safety Check' and it will tell you show you Chrome's safety settings.

You can adjust permissions under the tab labeled Permissions, which will take you to 'Site Settings.' Tweak to your heart's desire, everything from camera access to third-party cookie access.

More from Tom's Guide

• I've ditched Google for Perplexity — here's four reasons why
• I put the iPhone 17 vs Pixel 10 through a 7-round face-off — here’s the winner
• Employees are unknowingly leaking company secrets through ChatGPT, new report warns

Like many other services lately, Google is working on testing automated switching from passwords to passkeys in Chrome. Windows Report first spotted the flag hidden in the Canary build of Chrome which can automatically convert saved passwords into passkeys when a user is logged into a site or service.

The process doesn’t require any prompts or interaction from the user but will run automatically in the background once enabled. When you access a website and log in using a saved username and password, a passkey will automatically be created as long as the site in question supports it. This is supposed to happen invisibly, making the switchover easy for the user. Currently, the system in Chrome prompts users to adopt passkeys instead of passwords and requires confirmation after the prompt.

Passkeys are much safer than passwords because passwords can be easily guessed via either phishing or brute force attacks, are stored on third-party servers and are often reused or created in an insecure fashion in the first place. However, if you're not quite ready to switch to passkeys and plan to continue using passwords, make sure that you're using one of the best password managers to keep them as secure as possible.

How to turn on passkeys in Chrome

In order to turn on this new automatic passkey upgrade feature in Chrome, you first need to be running the experimental version of Google's browser, Chrome Canary. To do so, head to the Chrome Canary site and then download and install it on your computer.

It's worth noting that Canary is Google's most experimental and unstable version of Chrome, so you might run into some hiccups and other minor issues. However, it could be worth it if you want to test out the search giant's bleeding-edge features first.

With that out of the way, you then need to go to Chrome's Experiments page in your browser. This is done by typing "chrome://flags" into your address bar if you want to see all of the experimental features available in Chrome Canary. However, if you want to get right to the chase and start having Chrome convert your passwords to passkeys automatically, you can just enter "chrome://flags/#web-authentication-passkey-upgrade" instead.

Once you find this new experimental feature (either by searching for it or going directly to it with the URL above), you'll need to toggle it on. From there, restart Chrome Canary and then head to Google Password Manager to enable automatic passkey upgrades. And that's it, sites that support passkeys that you log into using Google Password Manager to autofill your credentials will now automatically convert your passwords to passkeys with Chrome's help.

If you don't want to go through the trouble of setting up Chrome Canary, then you'll just need to wait for this experimental feature to come to a stable version of Chrome. As big passkey enthusiasts ourselves, we'll be looking out for this change and will update this story once this useful new feature becomes more widely available.

More from Tom's Guide

• New hacker tool can inject AI-generated deepfakes right into your iPhone — everything you need to know
• Nearly 200,000 people hit in New York Blood Center data breach — names, IDs, SSNs and more exposed
• Google just fixed a major Chrome zero-day flaw — update your browser right now

Google just gave Chrome a serious AI upgrade, rolling out 10 new features powered by its Gemini model. These new features are designed to make your browser more productive, smarter and a lot more personal.

If you're like me and have 20 tabs open at any given time, and you've forgotten which page you saw those great running shoes on, Chrome’s new AI features are designed for you.

One feature can recall webpages you’ve visited, even if you can’t remember the name. Another actually summarizes entire tabs and answers questions on the fly. There’s even a “personal assistant mode” coming soon that can help you book appointments or handle errands online for you.

Out of the ten new features, these three are the ones I'm most excited about.

1. Memory for everything I've browsed

Starting this week for users in the U.S., Chrome now includes Gemini, Google’s AI assistant. Directly inside your browser, you can ask it to explain, summarize and compare all your tabs.

So next time you’re reading an article and want to dive into the details or get clarification, Gemini can break it all down for you. Or, if you’re comparing flights and hotel options, Gemini will give you a summary so you don’t have to click back and forth.

I can’t tell you how many times I’ve gone down a rabbit hole, closed my laptop, and then a week later thought: where was that thing I wanted to check out again?

With Chrome’s new AI memory, I can just type in natural language. Something like “What was that standing desk I saw last week?” and Chrome pulls it up instantly.

As someone juggling a lot, this one is sure to make my busy life much easier and productive.

2. Multitab AI magic

If you’re someone who leaves 15 tabs open while researching (guilty!), Gemini can now analyze them all together. That means accessing summaries and takeaways faster and less tab hoarding.

So instead of hopping back and forth between articles, I can get a single digest of what’s important. For someone who’s juggling work, researching, shopping for the family and keeping kids' schedules organized, this is a lifesaver.

3. Built in scam protection with Gemini Nano

This one feels less flashy, but honestly, it’s huge for me. Chrome now uses Gemini Nano in the background to detect phishing pages, fake virus pop-ups, and, yes, even those too-good-to-be-true giveaway scams.

I’ve seen enough shady pop-ups on my parents’ computer to know this is going to save a lot of people's headaches, and maybe even money.

Hopefully this also means fewer interruptions from pop-ups. You know the ones that suddenly appear the minute you start reading? Chrome now lets you automatically say goodbye to those, keeping them from interrupting you for a better user experience.

Bonus favorite: One-click password resets

Now, if one of your saved passwords gets exposed in a data breach, Chrome will offer a one-click way to reset it on supported sites (like Spotify, Duolingo, etc.). Fast, without disrupting your online experience.

The takeaway

More from Tom's Guide

• People are ditching ChatGPT for Gemini 2.5 Pro — here’s why
• This new AI tool can predict your risk of cancer and 1,000+ diseases — meet Delphi-2M
• Facial recognition AI meets live music — how Massive Attack made surveillance impossible to ignore

Chrome is getting a massive AI update in the coming weeks thanks to an infusion of new Gemini-based features making AI integral to nearly every aspect of using Google's browser. Google outlined the updates in a blog post today (September 18).

For those who've bought into AI, the new tools look useful; if you haven't, there's Vivaldi, a browser that is stridently anti-ai and built for online security (we would have suggested Firefox but it is also getting an AI chatbot).

You'll know Gemini has arrived thanks to a new Gemini button on the desktop browser. Here's a closer look at what this Gemini infusion brings to Chrome. The new features can be boiled down to three categories: online security, Google search, and tab/website organization.

10 new Gemini in Chrome features

• New Gemini button for enhanced browsing
• Agentic browsing assistant to complete tedious tasks
• Multitab organization and AI summaries
• AI prompts to find previously visited sites
• Integration with other Google apps like YouTube and Calendar
• Search AI mode right in the address bar
• Ask questions about content on a web page
• Enhanced Gemini Nano to stop scams and spam
• Block unwanted notifications and personalized permissions
• Change compromised passwords with one click

Gemini Assistant and Search

A new Gemini button will open up side tab where you can ask questions about content in your open tabs. On mobile, Gemini already works with Chrome, but in iOS it will be built into the Chrome app.

Google says that agentic capabilities are coming over the next few months to "handle tedious tasks" like booking a haircut or buying your food. You will tell Gemini what you want to do and it will act on web pages for you.

For those who have kept AI Mode turned on, you'll now be able to access it right in the address bar (or what Google has dubbed the omnibox). This mode is supposed to let you ask longer and more complex questions to receive AI responses.

In AI Mode, you can ask follow-up questions. This specific feature is supposed to launch before the end of September.

In normal search, you'll still get the AI Overview and you can follow up with questions about the content in the tab in AI Mode. There will also be contextual suggestions if you can't think of your own.

Tab organization

Google claims Gemini will be able to work across multiple tabs, meaning you could be doing trip research, while planning a flight and booking a hotel. Gemini is supposed to be able to consolidate all of that info into an itinerary if needed.

Or you could wait for the experimental Split Tabs feature to become a full-fledged release.

Additionally, Gemini is getting the ability to read your website history so you can give it prompts like “what was that blog I read on back to school shopping?" to find sites you previously visited.

Non-Chrome apps like Maps, YouTube and Calendar are getting Gemini integration that would let you schedule appointments, or look up a location without leaving the webpage you're on. Doing so involves asking Gemini in Chrome using the Gemini button.

Online security

A version of Gemini Nano is coming to Chrome as part of this AI expansion that would enhance the already existing Safe Browsing Enhanced Protection mode, which already uses Gemini Nano.

This mode IDs tech support scams to stop you from downloading potentially harmful software. The updated version is also supposed to stop malicious sites from hitting you with fake viruses or fake giveways to trick people into clicking.

Chrome will now use AI to learn your preferences to determine which sites are allowed to give you notifications while also blocking spam or scam notifications.

Lastly, Chrome will be getting an AI password agent that will let you change saved passwords with a single click, but only on supported sites. Examples provided by Google include Spotify, Duolingo and H&M. It's unclear how many websites will support this feature.

This is meant to expand on the password features that already exist in Chrome such as the ability to alert you about compromised passwords and login credentials autofilling.

When is this all coming?

Chrome is one of the most popular browsers in the world, so this Gemini integration is going to put AI features in front of more people in the coming weeks and months.

Not every Gemini feature announced by Google will be available right away. Some, as mentioned, are rolling out in the coming months while others like the updated AI mode are supposed to launch by the end of the month. Subscribers to Google AI Pro and Google AI Ultra will have already seen some of these features.

As with most Google rollouts, just because a feature launched today, it doesn't mean you'll get it right away. It could take a few days for you to see any of these updates in Chrome.

More from Tom's Guide

• Google just fixed a major Chrome zero-day flaw — update your browser right now
• Google Pixel’s new accessory is what the iPhone 17's Crossbody Strap should have been
• Android 16 just gave Google Maps a small makeover — and it's all thanks to Material 3 Expressive

For the sixth time this year, Google has issued a patch for a zero-day flaw in Chrome. According to Bleeping Computer, the emergency security update fixes a flaw that is currently being exploited by hackers in the wild.

As per usual, Google is not issuing any further details about the vulnerability at this time: “Access to bug details and links may be kept restricted until a majority of users are updated with a fix… We will also retain restrictions if the bug exists in a third party library that other projects similarly depend on, but haven’t yet fixed.”

The company did warn that this is a public exploit, which typically indicates that there has been active exploitation of the bug in the wild. In a security advisory released on Wednesday, Google stated that it is aware that an exploit for CVE-2025-10585 exists in the wild. The vulnerability is a high-severity zero-day caused “by a type confusion weakness in the web browsers V8 JavaScript engine” according to Google’s Threat Analysis Group.

The company mitigated the issue a day later by sending out a release for both the Windows/Mac and Linux versions of its browser. This version will roll out to Chrome's Stable Desktop channel in the coming weeks.

How to stay safe online

• Nearly 200,000 people hit in New York Blood Center data breach — names, IDs, SSNs and more exposed
• Samsung security flaw could let hackers remotely control your device — update your Galaxy phone right now
• Google just took down 224 malicious apps with 38 million installs in massive SlopAds fraud campaign — how to stay safe

When you have as many browser tabs and windows open at a time as I do, you’re always looking for new ways to organize them all and to be more productive at the same time. Well, I think I might have found my favorite one yet, and the best part is that you can enable it in Chrome right now.

You see, while others use Microsoft Office and other apps for their work, the majority of my research, writing and editing is done in Chrome as I’m a Google Workspace power user. This approach might not be for everyone, but given that I’m constantly switching from one computer to another while reviewing the best mini PCs, I love how once I sign into Chrome, my bookmarks and the rest of my work are right there waiting for me.

At any given time, I have at least three Chrome windows open, as I have two side by side on my ultrawide monitor up top and another for my work chats in Slack open on my portable monitor underneath it. I’ve gotten quite used to this setup over the years, but after trying Google’s latest experimental Chrome feature, it’s completely changed how I think about both browser windows and tabs.

Here’s everything you need to know about Chrome’s new Split Tabs feature, along with how it’s changed the way I use my browser and all the steps to enable it, so you can try it out for yourself.

What are Split Tabs?

From vertical tabs to tab groups and tab stacking, the big browser makers like Google, Mozilla and Microsoft, as well as the smaller ones like Vivaldi and Brave, have all been trying new things with how we view, organize and manage tabs.

For the most part, innovations like vertical tabs and tab groups didn’t really change my workflow. However, the same can’t be said for Split Tabs, as once I enabled this experimental feature, I immediately saw its potential.

I found out about them thanks to a new report from Chrome Unboxed. At first, though, I thought this was a Chromebook-only feature as that site does a fantastic job covering all the latest developments with ChromeOS and the best Chromebooks. In fact, a review from Chrome Unboxed was what convinced me to pick up a Lenovo Chromebook Duet 3 back when I first started at Tom’s Guide. Much to my surprise, when I opened the Chrome Flags menu on my Windows PC, I found the experimental Split Tabs feature right there and knew I had to try it out for myself.

With Split Tabs enabled, when you have two tabs open in the same Chrome window and right-click on one of them, a new option appears that reads “Add tab to new split view.” Clicking on this will take those two tabs and merge them into one combined tab. Basically, it appears as if you have two browser windows open side by side, but they're grouped together within the same tab.

From there, you can switch which website appears on the left or right side, resize the split panes to have one site appear larger than the other, separate the combined tab back into two traditional tabs and more.

One of my favorite parts of this new experimental Chrome feature is that within a Split Tab, you can have one site open on the left side and then open links from that site and have them appear on the right side or vice versa. Whether you’re comparing two sites to one another or doing the kind of research that would normally result in a dozen or so open tabs, Split Tabs are a game-changer in the truest sense of the phrase.

As part of its trial rollout of Split Tabs, Google has added a new menu icon to the right of Chrome’s address bar that lets you quickly swap which side a particular webpage is on. Likewise, you can also close one side or the other or separate a combined tab back into two standard tabs. And for those who don’t want to move their mouse up, you can access this exact same menu between your two Split Tabs at the bottom of a Chrome window, too.

There’s already a whole lot to like with this new experimental feature, and hopefully, Google adds even more to it once it rolls out as part of an official Chrome release. For now, though, here’s how to enable Split Tabs and try them out for yourself.

How to enable Split Tabs in Chrome

If you want to give Split Tabs or any other of Chrome’s experimental features a try, it’s really easy to enable them. To get started, head to your browser’s address bar and type the following: chrome://flags. This brings up Chrome’s Experiments menu, where, under a general warning about the risks associated with enabling them, there’s a full list of all of the current Experiments you can try out.

Given that there are hundreds of new ones to try out at any time, it’s a lot easier to use the search bar at the top of the page to find the particular one you’re looking for. With Split Tabs, for instance, you want to search for #side-by-side to bring it up. After toggling disabled to enabled on the right side of this Experiment’s description, all you have to do is restart your browser to test it out.

With two tabs open in a Chrome window, you’ll now have the option to merge them together into a single, combined tab. If you’re using an ultrawide monitor like myself, you can have two windows side by side, each with its own Split Tab. This almost makes it feel like you’re using a quad-monitor setup with four separate, full-sized sites visible at the same time.

Whether you’re doing research for a school project, planning a trip or even just shopping online, Split Tabs provides a whole new way to browse that, as a long-time Chrome user, has really impressed me so far. Hopefully, Google decided to make this experiment a full-fledged Chrome feature.

So, are you planning to enable Split Tabs and start using them? Let me know if there are any interesting ways you’ve incorporated Split Tabs into your own workflow or even just how you browse the web in the comments below.

More from Tom's Guide

• I tried a foldable triple-display portable monitor — and it’s the ultimate mobile accessory for power users
• Forget Chrome — this new AI browser is changing how people search the web
• I spent a week with this AI-powered mini PC and the buttons on the front earned it a permanent spot on my desk

Just over a year after a federal judge declared that Google is a monopoly in a landmark antitrust case, penalties have been handed down, though Google has been spared the breakup that the government had been seeking.

There was some expectation that the ruling handed down today (September 2) would force Google to divest various parts of its business including selling Chrome or separating Android from Google. That's not what happened, though Judge Amit P. Mehta of the U.S. District Court for the District of Columbia did impose some penalties.

As part of the ruling, Google must share some of its search data with companies that are "qualified competitors." Specifically, Google has to share search index and user interaction data, though it doesn't have to share advertising data.

Mehta also called for the creation of a technological oversight committee that is supposed to monitor Google's compliance with the ruling for the next six years.

Additionally, Mehta restricted Google's ability to enter into exclusive contracts for distributing its search engine. The company is still able to pay for distribution of its search and AI products. This means that Google could still pay $20 billion a year to Apple for making Google Search the provider on Safari, as long as the deal isn't exclusive.

The Department of Justice argued that Google should be forced to sell its Chrome web browser as a way to remedy the company's power as a monopoly.

“Notwithstanding this power, courts must approach the task of crafting remedies with a healthy dose of humility,” said Judge Mehta in the decision. “This court has done so.”

Mehta noted in his ruling that AI had upended the tech world since the lawsuit launched in 2020 and that factored in his decision.

Gabriel Weinberg, founder and CEO of browser maker DuckDuckGo, said that the ruling did not go far enough to force the changes needed to Google's behavior. "Google will still be allowed to continue to use its monopoly to hold back competitors, including in AI search," Weinberg said in a statement received by Tom's Guide. "As a result, consumers will continue to suffer. We believe Congress should now step in to swiftly make Google do the thing it fears the most: compete on a level playing field."

How did we get here

This is not necessarily the end of the lawsuit, as Google could appeal the ruling. For now, the landmark case is one of the first major attempts to tackle technology monopolies since US v Microsoft in 1998.

The Department of Justice filed the antitrust lawsuit against Google in 2020 arguing that the exclusive agreements with companies like Apple and Samsung was monopolistic and kept out competitors.

Mehta ruled that Google was a monopoly in 2024 with the case entering the remedies phase this past April. During this phase, Google agreed to halt the exclusive contracts but fought back on every other point including the forced sale of Chrome and sharing data.

Google argued that doing so would harm the company, reduce innovation and basically act as a free handout to competitors. However, as of this writing, even with the rise of OpenAI and its improved AI-based search, Google has nearly 90% of the search engine market according to GS Statcounter.

More from Tom's Guide

• I tested the Pixel 10 Pro — 5 reasons to buy (and 3 reasons to skip)
• Google Gemini’s AI Photo Editing Is Next-Level — Here’s What It Can Really Do
• Best cheap phones tested — Our top budget phone picks

A VPN is a common and recommended privacy tool when going online which is why many people download them as either apps for their computer or smartphone or as extensions for their browser of choice. However, not every VPN is trustworthy or even remains that way. Case in point, FreeVPN.One is a Chrome extension with 100,000 installs, a 3.8 star rating and even a verified badge. However, several recent updates have changed the status of the extension from safe to very suspicious.

As reported by Cybernews, researchers at Koi Security recently put out a report detailing their findings on the extension complete with a timeline that showed how it has shifted to spying on its users instead of protecting their privacy. This was done by taking screenshots of their displays and sending this data back to remote servers. When a webpage is loading, the extension will grab a screenshot in the background and then send it back to a server with additional details like the URL, tab ID number and a unique user identifier.

The report points out how screenshots can include sensitive and personal information like passwords, personal messages and banking details; sending these images to a third-party server is the exact opposite of what a privacy tool like a VPN is supposed to do for its users.

Additionally, the extension now collects and transmits other sensitive information like IP geolocation, device information and sends it encoded in Base64, or via AES-256-GCM encryption with an RSA key which hides the data in transit.

One of the features of the extension is “Scan with AI Threat Detection” which is intended to upload screenshots and URLs when the user clicks a ‘check URL’ button, however, there is no indication to the user that the extension has been repeatedly grabbing screenshots in the background even when this feature is not being used.

These suspicious features were added recently and went live on July 17 of this year, which was right before strict age verification rules went into effect in the U.K.; many people in that area would be searching for VPNs after the rules went into place as a workaround.

VPN extensions need some permissions to operate on a user’s system – usually proxy and storage permissions. However, FreeVPN.One required suspiciously more access including all URLs, tabs and scripting permissions. The researchers at Koi Security pointed out that this, in addition to the other issues the extension exhibited, opened the door to persistent surveillance of its users.

A developer is not named in either the privacy policy or the Terms of Service on the FreeVPN.One website, and their explanations to the security researchers about their findings do not match. According to the researchers: “The [developer] explained that the automatic screenshot capture is part of a Background Scanning feature, which should only trigger if a domain appears suspicious. In practice, we saw screenshots being captured on trusted services like Google Sheets and Google Photos, domains that cannot be considered suspicious” and the developer ceased communications when asked to provide proof of legitimacy.

If you've installed this VPN extension, you should uninstall it and run a scan using one of the best antivirus software solutions. Likewise, you're going to want to check your browser extensions to ensure that it's removed and then check your screenshots folder as well.

More from Tom's Guide

• Over 2 billion Gmail users at risk following database hack — how to stay safe
• FBI issues warning to all smartphone users — this dangerous new scam could be at your door
• 16 million PayPal accounts exposed on the dark web — emails, passwords and more for sale

AI startup Perplexity has made a jaw-dropping move: an unsolicited $34.5 billion all-cash offer to buy Google’s Chrome browser despite being valued at only around $14 billion, according to a report from The Wall Street Journal.

The timing is no accident. A U.S. court recently ruled Google holds an illegal monopoly in search, and one proposed remedy is forcing the company to sell Chrome. Perplexity’s bid positions it as the first major player ready to pounce if divestiture becomes reality.

According to the WSJ report, the company says it would keep Chrome’s open-source Chromium code, retain Google as the default search engine and invest $3 billion into the browser’s development over the next two years.

It’s also a signal that Perplexity, already developing its own AI-powered Comet browser, sees controlling a mainstream browser as key to winning the AI search race.

What is Perplexity?

Founded in 2022, Perplexity has quickly become one of the fastest-growing names in AI search. Its conversational answer engine combines live web results with generative AI to deliver concise, cited responses. Backed by Nvidia, SoftBank, and other major investors, the company has launched its own AI-driven browser, Comet, and is in talks to pre-install it on smartphones.

While still far smaller than Google, Perplexity’s rapid growth and investor backing give it the resources, and boldness, to make a play for one of the world’s most popular browsers.

What this could mean for you

• More competition: If Perplexity gains control, Chrome could see faster integration of AI-powered features.
• Big Tech in the hot seat: Regulators are taking a harder look at how browsers shape the search and AI markets.
• Comet isn’t the endgame: Perplexity’s bid underscores that browser dominance could be the next big AI battleground.

Bottom line

More from Tom's Guide

• I tested ChatGPT-5 vs Google Gemini 2.5 with 10 prompts — and there's a clear winner
• From hyper-personal assistants to mind-reading tech — this is how AI will transform everything by 2035
• Claude just got a major personalization upgrade — here’s what’s new

Google's Chrome 138 update killed uBlock Origin and other popular extensions, leaving millions without their favorite ad blocker. While Google claims this improves security, many users just want ad blocking back — especially when YouTube hits you with double ads and streaming services like Peacock force commercial breaks.

In this guide, we’ll show you two ways to restore uBlock Origin in Chrome — whether you already have it installed and need to re-enable it, or you’re starting fresh and want to install it from scratch.

Both methods involve enabling hidden Chrome flags that temporarily bring back support for older extensions. This workaround won’t last forever, as Google plans to remove these flags in future updates, but it buys you some time to enjoy ad-free browsing.

Here's how to re-enable or install uBlock Origin in Chrome.

How to re-enable uBlock Origin in Chrome

How to install uBlock Origin in Chrome

More from Tom's Guide

• I tried Gmail's amazing new tool — it makes cleaning up your inbox easy
• You can speed up your Wi-Fi by moving your mesh router to these spots
• Google Drive full? Here's how to get your storage back for free

Apple users should be sure to update their devices to iOS 18.6, particularly if they’re also Chrome users, as the latest software version includes a security patch that addresses a high-severity exploit that has been targeting the browser in zero-day attacks.

As reported by Bleeping Computer, the zero-day flaw in question (tracked as CVE-2025-6558) involves an incorrect validation of untrusted input in an open-source graphics abstraction layer.

It then processes GPU commands and translates API calls, which enables remote attackers to execute arbitrary code within the browser’s GPU process via specially crafted HTML pages. This could potentially allow them to escape the sandbox that isolates browser processes from the underlying operating system.

All this to say that, according to BGR, if you don’t update Chrome, you could be opening yourself up to being attacked just by visiting a malicious website. Attackers could run code on your device and bypass the protections that normally keep your browser safe and from there perform other malicious activities like running malware on your phone, stealing passwords, deploying ransomware or botnets or recording clipboard or webcam views.

The flaw was discovered in June by the Google TAG team and reported to Chrome, who patched it in July and tagged it as actively exploited. As per usual, Google has yet to provide much additional information about the attacks. However, it is well-known that TAG is often responsible for discovering flaws that are exploited by government-sponsored threat actors in targeted campaigns.

Apple released WebKit security updates on Tuesday that addressed the vulnerability, which affects iOS 18.6 and iPadOS 18.6, macOS Sequoia 15.6, iPadOS 17.7.9, tvOS 18.6, visionOS 2.6, watchOS 11.6.

In its security statement, Apple said that “Processing maliciously crafted web content may lead to an unexpected Safari crash.” CISA (Cybersecurity and Infrastructure Security Agency) added this security bug to its catalog of vulnerabilities known to be exploited in attacks on July 22, requiring federal agencies patch their software by August 12th.

Updating your phone, computer and of course, your browser, is something you should be doing regularly and a zero-day flaw like this one which could be exploited by hackers in their attacks is exactly the reason why.

More from Tom's Guide

• More than 250 malicious apps are spreading info-stealing malware on Android and iOS — delete these right now
• 12 signs your phone has been hacked — and what to do next
• Apple just announced 5 big upgrades to protect your kids online — here's what's coming

AI is changing how we interact with the internet, and it’s happening faster than most people realize.

As tools like ChatGPT, Claude and Gemini provide instant answers without requiring users to click through to websites, the ripple effects are starting to show.

I've pointed to a steep decline in traffic for major publishers because of advanced AI overviews and chatbots hopping paywalls, suggesting that the traditional, ad-supported model of the open web is under pressure.

The worry is that if no one’s clicking, then no one’s paying, and that puts the future of free, high-quality content at risk.

The rise of AI-powered browsers

At the same time, a new generation of AI browsers is starting to emerge. OpenAI, Perplexity, Opera and others are building browsers that go beyond simply displaying web pages.

These tools are designed to function more like intelligent assistants becaues they can summarize content, complete tasks and guide users through multi-step processes like booking a reservation or comparing products across tabs.

OpenAI is reportedly developing a browser based on Chromium with GPT-4o integration. It’s expected to offer task automation, tab management and AI-powered content summarization.

Perplexity’s Comet browser, currently available, takes a similar approach but is gated behind a premium subscription.

Opera’s AI browser leans into privacy, offering local-agent capabilities that aim to keep user data on-device while still providing intelligent support.

What this means for the web

These features can be helpful, especially for users looking to get things done more efficiently. But they also raise big questions.

If AI can provide an answer without sending monetizable traffic to the original source, what happens to the websites that created that content in the first place?

This represents a shift in the long-standing structure of the web. For years, the exchange was simple: users got access to free information, and publishers got paid through ads and affiliate links. If AI bypasses that interaction, the balance starts to fall apart.

Browsing is becoming conversational

The idea of “surfing the web” is being replaced by something more passive and curated.

Instead of typing a query and scanning a list of results, users might soon ask their AI browser a question and receive a synthesized answer pulled from multiple sources, often without visiting any of them.

These AI browsers are reframing the internet as a kind of workspace, where information is delivered on demand and most of the heavy lifting is handled in the background.

It’s a shift in how we find and engage with information, and it could change the way users experience the web entirely.

Privacy and regulation on the horizon

There are also growing concerns about privacy. While Opera has taken a more transparent, local approach, many AI tools still rely on cloud processing.

That opens the door to user data being collected, stored and potentially used to train future models.

As AI browsers become more capable and more integrated into our daily lives, the trade-offs between convenience and privacy are likely to get more complicated.

On top of that, regulatory questions are piling up. The U.S. Department of Justice is already looking into Google Chrome’s market dominance, and lawsuits around AI data scraping are picking up steam.

Some companies, like Perplexity, are experimenting with revenue-sharing models for publishers, but it’s not yet clear whether those efforts will be enough to support the kind of content people rely on.

What comes next

AI browsers are poised to change the way we use the internet and that includes how we access information, what we trust and how we support the sites we visit.

Whether this shift ultimately hurts or helps the web depends on how companies, publishers and users respond to the changes ahead.

More from Tom's Guide

• Sam Altman just teased GPT‑6 — and it's more personal than ever
• I just created 5 apps using GPT-5 and it's so easy it feels like it should be illegal — here's how to do it
• I just tested ChatGPT-5 vs Deepseek with 9 prompts — and there's a clear winner

Microsoft is at it again with another ploy to stop Bing users from switching over to Google Chrome, this time with a full banner right on the search page.

As Microsoft Edge is Windows' default browser, many users often use it to download Chrome through the Bing search engine. Many will know that the browser often delivers a message stating why it's worth using Bing instead, but Windows Report has noticed a new comparison banner pops up.

When typing in "Chrome" using Bing on Edge, there's now a banner that takes up a big part of the page before web search results are shown. The report found a comparison between the two browsers, stating " Microsoft Edge offers the speed of Chrome with the security of Microsoft."

Notably, it also details what Edge offers and what Chrome doesn't, including Microsoft Rewards, a built-in VPN (although this isn't available in all regions), AI personalization and Microsoft's "recommended" for optimized browsing.

This is a "Promoted by Microsoft" ad, and a large one at that. It's yet another way Microsoft is urging users to stay on Edge, although I'm not sure if this will stop users in their tracks.

When trying this myself, I received another banner that showed the Microsoft Reward points to gain by trying Edge, with tiles showcasing Amazon, Starbucks, Xbox Game Pass and Roblox gift cards.

Either way, instead of just a small pop-up notification, there's now a full-on banner before reaching the list of search results for Google Chrome.

With Chrome continuing to be the most popular browser around, it comes as no surprise that Microsoft wants you to stick with Edge. According to statistics from analytics firm Backlinko, there are an estimated 3.69 billion users worldwide, compared to Microsoft Edge's estimated 292 million users.

Thankfully, those in the EU won't have to put up with Microsoft begging you to make Edge your default browser, thanks to the Digital Markets Act. However, the rest of the world will still get pop-ups to give Edge a go.

Interestingly, research has shown that Chrome is the hungriest browser around, with Microsoft Edge using far fewer system resources. However, Bing itself, which Edge uses, is only a few steps behind Chrome in the amount of data it collects.

With Microsoft Edge now being promoted as an AI browser, it still has a lot of competition in the market, with Perplexity AI launching its first AI browser, and Opera offering its own agentic AI browser.

We're sure to see more tactics Microsoft uses to turn Edge into the browser for all (instead of Chrome), but if you want to uninstall Microsoft Edge instead, that's always an option.

More from Tom's Guide

• Android head says 'we’re going to be combining ChromeOS and Android into a single platform'
• Move over, OpenAI! Opera Neon is a new AI browser that can surf the web and even build games while you sleep
• How to stop your personal data from appearing in Google searches

My Google Drive is full, and I keep getting that annoying "storage full" notification. Before paying for a Google One subscription, I decided to see what I could do with my free 15GB.

Turns out, I had tons of unnecessary files taking up space that I didn't even know about. Between forgotten email attachments, duplicate photos, and cached data I never use, I was wasting gigabytes of storage.

I found several ways to free up space without spending anything. Some methods were obvious once I discovered them, others took some digging around Google's settings, but they all helped me reclaim some serious storage space.

Here's exactly what I did to clear out my Google account and extend my free storage life without paying a penny.

The truth about Google storage limits

Google gives you 15GB of free storage shared across Drive, Gmail, and Photos. This means a single large video file or years of accumulated email attachments can quickly consume your entire allowance.

The key to maximizing free storage lies in understanding that not all files are created equal. A few large files often take up more space than hundreds of smaller documents, making strategic deletion more effective than random cleanup efforts.

Now you've learned how you can increase your Google Drive storage space, why not take a look at some other useful guides?

Check out Google is tracking you right now — change these 3 settings to stop it and how to replace your PC GPU — 16 easy steps to upgrade your desktop quickly and easily.

Chrome OS is good and all, but it’s always felt distant — on an island of its own away from the might of Android. Well, it turns out Google finally got the memo to send a rescue boat out to said island, as an executive has confirmed the company’s plans to combine the two platforms.

This comes in an interview with TechRadar, where Sameer Samat, president of Android ecosystem at Google, just dropped this nugget into a conversation about why the interviewer was using a suite of Apple devices.

Now, we have heard this before. Back in November, Android Authority reported that a source inside Google had said this was the mission. However, this is the first time Google has confirmed it on-the-record. And honestly? It may be a few years late in my eyes, but I’m glad it’s finally happening. Let me explain.

Android’s hitting the big screen

This comes off the back of Android bringing some new big screen features to the table, such as a proper desktop mode a la Samsung Dex, improved adaptability of apps, better external display support and windowing within the desktop.

Basically, Android is now able to do pretty much everything that a Chromebook could, so why have the two running side by side in isolation of each other?

I’ll always appreciate what Chrome OS has brought to the table, though. I mean with the Lenovo Chromebook Plus 14, Google was able to make low-key one of the best laptops for students, sharpen up the OS with quick shortcuts and add AI features that are logically added rather than clumsily overlaid (looking at you, Microsoft Copilot).

But what was learnt in Chrome OS can also come over to Android too, and having it all in one place unlocks a special new perk on the computing skill tree.

‘It just works’

This is the moniker you hear around a lot of Apple devices. It’s one of the reasons why I use them, why Lance who conducted this interview at TechRadar uses them, and hell, it’s probably why a lot of you out there use them.

The fact that everything just comes together through Continuity features and marries these devices together nicely is always one of Apple’s superpowers that Google has been able to do somewhat, but just not to the level of smoothness you’d find in the Cupertino playground.

To do it properly, you need real sync up in the operating systems, which for all the heavy lifting Google did with Chrome OS to try and make it talk to Android, it was just not feasible. But now a proper merger is on the cards, this is the answer to all of its problems.

And stat-wise it makes sense too. According to most recent data, around 1.25% of the world’s computer users are on Chrome OS — down from nearly 2% last year. With a downward trajectory, change needed to happen, and this might just be the smartest decision the company’s made in a decade.

So what now?

As this was a quick quote from a Google exec, there are some questions I have that remain unanswered as of now:

• Does this mean any Android device could be a desktop device too? Imagine the Pixel 10 getting a laptop-esque accessory (like the long-gone NexDock) to browse your device like a laptop. Pretty cool, right?
• What does this mean for current Chromebooks? Android does technically work on Arm and the more computer-traditional x86 architectures (I go more into the differences in my review of the AMD-armed Asus Zenbook S16), but we’re talking over 99% of its features and apps have been built with Arm in mind. What happens to Chromebooks that are running older Intel chips? Are hardware requirements going to change drastically and leave these people in the cold?
• What about updates? Chrome OS is usually updated every four weeks, mostly for security reasons but Google does sneak some new features in there too. Will moving to Android’s more infrequent feature drops affect this?

We don't have answers to any of these, and it'll be just a case of waiting and seeing what happens (my guess is by the time Android 17 rolls around). But as far as the initial news itself goes, this is an encouraging step in the right direction for Chromebooks.

More from Tom's Guide

• These Prime Day laptop deals are somehow still live — this is your last chance before Black Friday
• Apple's entire Mac lineup until 2026 revealed in leak — here's every MacBook, Mac Pro and more to expect
• I review over 200 laptops each year: here are the best laptops of 2025

If you’ve ever had the creeping sense that Google knows a little too much about you, you’re not wrong. From your search habits to your exact location on any given day, Google keeps a detailed log of your digital life, all in the name of personalization.

But convenience comes at a cost. Whether you’re concerned about data privacy, targeted ads, or just want more control over your information, the good news is you can push back.

It only takes a few setting changes to limit what Google collects, what it remembers, and how it uses your activity. Here’s how I adjusted my Google account to take back a little privacy.

Now you've learned how to stop Google tracking you, why not take a look at our other useful guides?

To protect your privacy even further, check out how to stop your personal data from appearing in Google searches and how to check for hidden cameras in your Airbnb and hotel.

And if you want to safeguard your phone when entering the U.S., don't miss these 5 security tips.

Almost a dozen malicious extensions found in Google’s Chrome Web Store have been downloaded 1.7 million times. According to Bleeping Computer, these unsafe add-ons are still largely able to act as legitimate tools but meanwhile they’re also able to to track browser activity, track users, and potentially redirect users to web addresses that could spread malware.

Discovered by researchers at Koi Security, a company that provides a platform for security self-provisioned software, the extensions in question range from VPNs and weather forecasters to themes and keyboards. The researchers also reported the extensions to Google; while some of them have been removed, some continue to be available.

The extensions in question are:

• Color Picker, Eyedropper
• Emoji keyboard online
• Free Weather Forecast
• Video Speed Controller — Video manager
• Unlock Discord — VPN Proxy to Unblock Discord Anywhere
• Dark Theme — Dark Reader for Chrome
• Volume Max — Ultimate Sound Booster
• Unblock TikTok — Seamless Access with One-Click Proxy
• Unlock YouTube VPN
• Unlock TikTok
• Weather

At least one of those, Volume Max – Ultimate Sound Booster – has been previously flagged by a different set of researchers who were concerned about its potential for spying on users.

Many of the extensions are verified, have hundreds of positive reviews and are prominently featured, which not only misleads users about their safety but also may indicate that these extensions were hijacked by threat actors who then introduced malicious code.

Because malicious code was introduced at a later time via updates and because Google’s auto update system will deploy the newest version to users without requiring user interaction, the code was rolled out to users without them knowing.

The Chrome Extensions API is used by the malicious extensions to execute their functionalities in the background, registering a listener that is triggered every time a user navigates to a new webpage.

This listener captures the URL of the new webpage and exfiltrates the information to a remote server with a tracking ID for each user; the remote server can respond with redirection URLs, which hijacks the user's browsing activity and could potentially take them to unsafe destinations. This could lead to cyberattacks (though is not something that Koi Security observed in their testing).

Koi Security also discovered similar behavior in the official store for Microsoft Edge, with a total of 600,000 downloads – combined, this creates one of the largest browser hijacking operations the researchers say they’ve ever documented.

What to do if you've downloaded the malicious extensions

First, remove any and all of the listed extensions from your machine. Then make sure to clear all of your browsing data to get rid of any trackers or tracking identifiers.

Next, check your system for malware by running a scan using your antivirus software. Keep an eye on your accounts and monitor them for any suspicious or unusual activity. Many antivirus programs have features that can help you keep track of your accounts, watch the dark web, or have features like identity monitoring.

More from Tom's Guide

• More than 1,200 fake Amazon sites pop up ahead of Prime Day — avoid getting scammed
• Best Prime Day antivirus deals: 7 heavily discounted security suites to keep you safe online
• This dangerous Mac malware just got a major upgrade which makes it even harder to delete — how to stay safe

Google has issued an emergency security update patch for Chrome in order to fix three security issues including one zero-day bug that has been actively exploited in the wild by hackers. This makes it the third active vulnerability to be patched via emergency update since the start of the year, with two others occurring in March and May.

As reported by Bleeping Computer, the latest flaw, tracked as CVE-2025-5419, is a high-severity vulnerability caused by an out-of-bounds read and write weakness in the V8 JavaScript and WebAssembly engines in Chrome.

It was initially reported on a week ago by members of Google’s Threat Analysis group; Google has confirmed that it is being exploited in the wild though the company is not sharing much additional information at the time as they are waiting until more users have had an opportunity to patch their browsers.

In the security advisory published on Monday, the company is quoted as stating: “Access to bug details and links may be kept restricted until a majority of users are updated with a fix. We will also retain restrictions if the bug exists in a third party library that other projects similarly depend on, but haven’t yet fixed.”

This is typical when it comes to active exploits, as it keeps other threat actors from hopping on the band wagon to take advantage of the vulnerability before users are able to update the fix. However, reporting from The HackerNews, says that the flaw involved allowing a remote attacker to potentially exploit heap corruption via a crafted HTML page.

Google reports that the issue was mitigated a day after it was reported via a configuration change that was pushed through the Stable Desktop channel across all the Chrome platforms. The zero-day flaw was likewise corrected the same day with updates to Chrome that are rolling out to users in the coming weeks.

Chrome does automatically update when new security patches become available, however users can make sure the installation is completed by going to the Chrome menu > Help > About Google Chrome.

Let the update finish then click Relaunch in order to make sure the patch has installed. The update versions are 137.0.7151.68/ .69 for Windows and macOS and version 137.0.7151.68 for Linux. Users of other Chromium-based browsers (Edge, Brave, Opera, Vivaldi) should apply the updates as they become available.

More from Tom's Guide

• Hackers are using this to spread dangerous malware just in time for summer travel season
• Microsoft Authenticator will shut off the password autofill feature in July — here’s how to save them
• It’s time to stop believing these lies about antivirus software

For regular users of Chrome for Android, using two-factor authentication on various websites requires you to open your messages app to retrieve verification codes sent via SMS.

In general, most Android apps are already friendly with the codes as they can detect them and automatically fill them in.

Currently, in Chrome, you need to copy or type the code into the website manually.

That might change soon, though. Over on Reddit, tipster Leopeva64 (via Android Authority) found that Chrome for Android might soon automatically detect the two-factor authentication codes and then fill them in.

The auto detect and fill feature saves time in the apps and would be appreciated by people who prefer visiting websites in Chrome, especially those that don't have apps.

"And yes, many users consider this verification option insecure, but several sites still use it, so Google wants to make the process of filling out these codes easier," Leopeva64 added in a comment on their post.

They noted that the feature is available in the "Canary version," but has not yet been turned on.

While the welcome feature isn't out (and may not be released), this week, Google revealed some new Gemini-powered accessibility updates, including expressive captions and an optical character recognition tool that helps screen readers recognize PDFs.

More from Tom's Guide

• Google is taking Gemini beyond smartphones — here’s what’s coming
• Google just fixed a high-severity Chrome flaw that can be used to take over your account — update right now
• Google's adding more accessibility features to Chrome and Android — and they're powered by Gemini

The internet is designed to collect your data. Whether it's cookies, trackers, or permissions, your data is always up for grabs – and your browser is one of the most guilty suspects.

One of the best VPNs, Surfshark, has analyzed 10 of the most popular mobile internet browsers to determine how much data each one collects.

Google Chrome collects the most data by far, and 30% of the browsers analyzed share your data with third-parties.

Safari also collects a large amount of data, and together with Chrome, the two occupy 90% of the mobile browser market share. That's a lot of data.

But not every browser analyzed collects huge amounts of data, and there are steps that can be taken to avoid mass data collection – here's what you need to know.

This article was edited on May 19, 2025, to clarify "coarse" location data and its implications.

20 different data types

Surfshark's research found that Google Chrome was the most data-hungry browser, collecting 20 different data types.

Contact information, financial details, location, browsing history, usage data, and identifiers are just some of the data Chrome collects. It is thanks in part to the Google ecosystem, with Google Drive, Maps, and Mail all seamlessly integrated into the browser.

Chrome is the only browser analyzed that collects financial information such as payment methods, card numbers, and bank account details. It's also the only one to collect contacts from the user's phone or address book. However, it is worth noting that some of this information is provided voluntarily by users – having Chrome remember card details for example.

Bing was the second-hungriest browser, collecting 12 data types. Pi Browser was third, collecting nine types of data. Safari and Firefox were tied for fourth place, each collecting eight data types.

Some browsers are better than others

The average number of data types collected across the 10 browsers was six, and only two browsers collected fewer data types than this.

The privacy-focused browser, Brave, collected only two data types – identifiers and usage data.

The Tor Browser was by far the most secure browser analyzed, collecting no data. Tor is a decentralized network designed to protect user anonymity and prevent tracking. It's certainly not the fastest or feature-packed browser, but is an excellent choice for maximizing privacy and security.

DuckDuckGo and Firefox are two browsers that have decent privacy credentials, and Surfshark rated their data collection as "moderate" – Opera also falls into this category.

They avoid collecting sensitive data, but do collect contact information, identifiers, usage data, and diagnostics.

They could be seen as browsers for those with moderate privacy concerns who still require robust browsing capabilities.

Third-party data sharing

30% of the browsers analyzed shared data with third-parties. Data is collected for advertising purposes, with Opera, Bing, and Pi Browser being able to do so.

40% of browsers were also found to have the capability to collect your location. However, Surfshark's data needs a little clarification, here.

Bing was the only browser to collect precise location data, which is a big red flag.

Safari, Chrome, and Opera are stated to collect "coarse" location data.

In human terms, collecting a user's "coarse" location means that the user can give consent to the browser to know their location, but it is not necessarily active by default.

This does not mean that these browsers are tracking your every move, and they require you to accept these terms before any location data is stored.

Given Opera's reputation as a privacy-focused browser, we asked them about these findings.

Michael Tegos, Product Marketing Manager, Privacy and Security, at Opera said it is open about its data collection and does so in order to "make better browsers" and monetize its free products. It details the data collected in its privacy statement.

"We do this only upon consent and in such a way that our users remain anonymous to us, through the use of randomized user IDs that are not associated with any real-world identity," Tegos added.

Opera users can revoke this consent at any time through their browser settings, and all data collection can be opted out of – which is more than can be said for many other browsers.

Tegos also highlighted Opera's built-in Ad Blocker and Tracker Blocker, as well as its built-in VPN.

"Opera VPN is also a no-log service, meaning that it never collects or stores user data," said Tegos.

"Our no-log policy has been independently audited by Deloitte, a trusted auditor of VPN services. We also publish a bi-annual Transparency Report, outlining how many requests for data we get from law enforcement authorities."

First steps to take

Tomas Stamulis, Chief Security Officer at Surfshark, shared initial steps users can take to secure their personal data.

• Allow permissions: Storage (to save downloads), a camera and microphone if you use video calls.
• Deny or restrict permissions: Location, contacts, SMS, and phone access – unless absolutely required.
• On Android: Use “Only While Using App” or “Ask Every Time” when granting permissions like location or microphone.
• Choose privacy-focused browsers that collect data strictly necessary for browsing.
• Turn off the personalization settings on your browser app: Syncing with your Google/Microsoft account, Ad personalization, Activity tracking (location history, web & app activity)
• On Android: Go to Settings > Privacy > Ads > Opt out of ad personalization
• Revisit app permissions monthly – many apps request more than they need over time.

"In public, it is often explained that browsers collect the data to personalize and improve a user's experience," Stamulis said.

"If we rely on this explanation, it appears that the more data you provide, the smoother and seamless your browsing should be. In any case, this study from Surfshark is a reminder to look through your browser app permissions," he added.

Use a VPN and a tracker blocker

In order to protect your data when browsing, you might want to consider using a VPN. Your data and internet traffic is encrypted, plus your IP address is masked.

Leading, reputable VPN providers will not log or share your data, keeping it safe from hackers and third-parties. Verified no-logs policies back these claims up, and Windscribe's recent legal case proved just how important they can be.

Some VPNs offer tracker, ad, and malware protection as part of their plans. NordVPN's Threat Protection Pro is a good example of this, and has been named as the best VPN malware protection.

But VPNs alone can't protect you from every single piece of data collection or trackers. Therefore, using a dedicated tracker blocker alongside a VPN is a great way to ensure your data is not tracked.

These products often include an ad-blocker as well, with Ghostery and uBlock Origin being some popular choices.

Google Chrome has recently U-turned on its use of trackers. Considering its global popularity, there will be many people who could benefit from tracker protection.

Switch to a secure browser

Surfshark's research shows Brave is a safe choice for those seeking a more familiar browsing experience. Its design is similar to Chrome and it's packed full of features. Despite having the capability of collecting a little more data, Opera is also a good choice for those looking to increase privacy due to its clear, consent-based model.

For total anonymity, the Tor Browser is best. However, its speeds are not the fastest and you won't get nearly as many features as mainstream browsers.

Vivaldi is a browser Surfshark didn't analyze but it's another solid privacy-focused choice. Vivaldi has also partnered with Proton VPN, and Proton VPN Free is now offered as a Vivaldi browser extension.

Mullvad VPN has built its own browser, in partnership with the Tor Project. Mullvad is one of the most private VPNs and its browser is no different. It's also open source and free to use – whether you're a Mullvad VPN user or not.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

Google has U-turned on its plans to roll-out a standalone prompt for third-party cookies and trackers in Chrome, leaving the data privacy of millions of users at risk.

In a blog post, Google said: "We’ve made the decision to maintain our current approach to offering users third-party cookie choice in Chrome, and will not be rolling out a new standalone prompt for third-party cookies."

Unlike privacy software like many of the best VPNs, major tech companies don't care about protecting the personal data of their users and have questionable data privacy practices.

Trackers, targeted ads, and cookies are all part of an ecosystem designed to profit and function from the collection of our personal data. But there are ways to combat this and wrestle back control of our data.

Use a tracker blocker

A tracker blocker is software designed to prevent cookies from tracking you across the internet. They often come in the form of browser extensions and are an excellent way of protecting your privacy.

The non-profit digital rights organization, the Electronic Frontier Foundation (EFF), has a free browser extension known as Privacy Badger. It prioritizes tracker blocking rather than ad blocking, only blocking ads that actively track you, and automatically learns to block invisible trackers.

Ghostery is another popular and free browser extension. The ad and tracker-blocker prevents private data from leaving your browser and regularly updates its tracker database.

uBlock Origin is also a popular ad and tracker-blocker, but it's currently unavailable on the Google Chrome Store. However, it can be downloaded via GitHub.

Some leading VPNs include tracker blockers in their plans. NordVPN and its Threat Protection Pro feature is a good example. It is a threat protection feature that blocks ads and trackers – it is also the best VPN malware protection.

Threat Protection Pro has already blocked over 177 billion trackers in the first three months of 2025, as well as 5 billion ads and 960 million cases of malware.

As well as NordVPN, ExpressVPN and Proton VPN both provide the ability to block trackers in their VPN settings but NordVPN's feature is the standout.

VPNs mask your IP address which can prevent you from certain, IP-related trackers. But using a VPN alongside a dedicated tracker blocker, or one that is bundled with your VPN plan, is the safest solution.

Switch to a secure browser

Major browsers are packed with trackers as standard. There are tracker-preventing settings and incognito modes in many, but if you want to truly protect your online privacy, then consider switching to a secure browser.

Popular browsers include Brave, Opera, Firefox, and DuckDuckGo, and all are designed to prevent trackers as standard.

Many also protect against "fingerprinting." This is when a list of user characteristics, including your device hardware, browser, and website location, is collected.

Seemingly insignificant data such as installed fonts or screen resolution can also be collected and stitching all this information together can create a unique and detailed "fingerprint" of your device.

Some secure browsers have built-in VPN protection or offer one as a paid add-on. Secure browser Vivaldi has recently partnered with Proton VPN to offer Proton VPN Free as a browser extension.

The EFF has a "Cover Your Tracks" tool which allows you to test your browser for tracking protections.

One of the most private VPNs, Mullvad, has developed its own secure browser in collaboration with the Tor Project. Private browsing is enabled by default, fingerprinting is protected against, and uBlock Origin is used to block ads.

Investigate and adjust settings

Big tech companies and major browsers benefit from tracking you, but you do have some control over the collection of your data.

You can limit some data collection and how much information is shared. Check your browser settings and disable as much tracking and data collection as possible. Using incognito mode as standard and deleting your browsing history are other actions you can take.

Chrome users can decide how ads are targeted through its "My Ad Centre" tool. It's worth remembering that Google's ecosystem – including YouTube and Gmail – can also track you and share data across profiles and devices, so change your settings for all accounts.

Although Chrome isn't removing all of its privacy protections, it's still one of the worst offending browsers for data collection and employees digital fingerprinting.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

With the release of Chrome 136, Google will finally address a long-standing privacy issue in its browser that allowed websites to determine your browsing history by using previously visited links.

Have you ever done a Google search for something and noticed that some of the links in the search results were purple? Well, this is because you previously visited those sites. As BleepingComputer points out though, those links are stylized as “:visited”.

If these purple links were only visible to you, this problem wouldn’t exist. However, since this color change is displayed regardless of which site you were on when you clicked a link, other websites could use creative scripts to leak your browsing history.

Over the years, security researchers have shown how this privacy issue could be used by hackers in a number of different attacks. Fortunately though, Google plans to finally address this 20-year problem with the release of Chrome 136 at the end of this month.

In the next version of Chrome, Google will use triple-key partitioning with “visited” links. That way, instead of storing link visits globally, the browser will partition visited links by link URL, top-level site and frame origin.

Once this change rolls out, it will ensure that a link will only appear as “:visited” on the same site and in the same frame origin. This will eliminate cross-site history links while protecting your privacy.

How to fix this privacy issue right now

If you don’t want to wait until the end of the month, there’s a way you can fix this issue now using Chrome Flags.

For those running, Chrome versions 132 to 135 (latest) you can enable the experimental feature :visited isolation right now. To do so, either head to chrome://flags and search for “Partition the Visited Link Database, including ‘self-links’” or you can just head to the following address: chrome://flags/#partition-visited-link-database-with-self-links.

Once there, you’ll see that the feature is turned off by default but if you click on the box on the right side, you can turn it on by selecting Enabled.

It’s worth noting that as this is an experimental feature, it might now work as expected in all situations. However, at least you have the option to address this privacy issue early.

For extra privacy though, you might also want to consider using one of the best VPN services while surfing the web since all of your web traffic and your searches go through a VPN server and can’t easily be seen by snooping websites and other third parties.

More from Tom's Guide

• Chrome extensions with 4 million installs are putting your privacy at risk
• Microsoft just patched 134 Windows security flaws including a zero-day used by hackers — update your PC right now
• Dangerous Windows malware can hijack your PC, hold it for ransom, steal your passwords and even let hackers spy on you

In an apparent reaction to the infamous Honey scandal that rocked the internet at the start of the year, Google has altered the rules around Chrome extensions to block the tactics the Honey extension is believed to have used.

Normally, a change to a Google affiliate ads policy wouldn't be newsworthy. But with the outrage from popular online creators over the potential loss of income from Honey's method of quietly taking credit for directing users to buy products, this is a big deal.

You can read the new rules in full on Google's updated Program Policies page (via The Verge), but here are the key parts:

"Affiliate links, codes, or cookies must only be included when the extension provides a direct and transparent user benefit related to the extension's core functionality. It is not permitted to inject affiliate links without related user action and without providing a tangible benefit to users. Some common violations include:

a. Inserting affiliate links when no discount, cashback, or donation is provided.

b. An extension that continuously injects affiliate links in the background without related user action."

Put simply, users of an extension now have to use or request an affiliate link via an extension knowingly. Also, that link must do something positive beyond just putting a company's tag on the purchase.

This is relevant to the Honey case since it's accused of secretly swapping affiliate links of other outlets with its own, and using inferior discount codes in order to claim more sales for itself.

Users must be aware

There's a further relevant section worth looking at here:

"Related user action is required before the inclusion of each affiliate code, link, or cookie. Some example violations include:

a. An extension that updates a shopping-related cookie without the user's knowledge while the user is browsing shopping sites.

b. An extension that appends an affiliate code to the URL or replaces an existing affiliate code in the URL without the user's explicit knowledge or related user action.

c. An extension that applies or replaces affiliate promo codes without the user's explicit knowledge or related user action."

In this section, Google states that extensions can't replace one affiliate link or promo code with another without the user's knowledge either, which is precisely another tactic the Honey extension is accused of doing.

Legal action against Honey is underway, so it's unclear how this story will end. But with Google's updated policies, it happily seems we're unlikely to get a sequel.

Honey claimed that it "follows industry rules and practices, including last-click attribution," in a comment to Tom's Guide when the story first broke, but it now seems those rules are changing to close up what could be considered an exploitable loophole.

More from Tom's Guide

• Chromecast fail — users across the world are saying that their Chromecasts are not working
• MacBook Air M4 benchmarks are here — here's how it compares to the MacBook Pro, Windows laptops and more
• iOS 18.4 has dropped — 5 new iPhone features to try first

Just like malicious apps on your smartphone, malicious browser extensions can put your devices and the sensitive data stored in your browser at serious risk.

The problem with malicious browser extensions is that, unlike with the apps on your phone, you never see them after installation unless you go into Chrome’s Manage Extensions menu. If these harmful extensions weren’t a big enough problem already, a new polymorphic attack allows them to instantly transform their appearance to mimic the legitimate ones installed in your browser.

As reported by BleepingComputer, the cybersecurity company SquareX Labs devised this new attack as a proof of concept. If it ended up in the hands of hackers, though, it could be really dangerous as malicious extensions could then be used to impersonate the best password managers and other extensions for banking apps, crypto wallets and more, which all store loads of sensitive personal and financial data.

Here’s everything you need to know about this new polymorphic attack and some tips and tricks on staying safe from malicious extensions after your browser data.

Fooling the Chrome Web Store

To pull off this attack, a hacker, scammer or other cybercriminal would first need to submit a malicious extension with polymorphic capabilities to the Chrome Web Store. The extension itself would need to do exactly what its listing page describes for it to bypass Google’s stringent security checks, though.

Once installed, this bad extension then abuses Chrome’s own ‘chrome.management’ API to get a list of all of the other extensions installed in a victim’s browser. If it doesn’t have permission to access this API though, SquareX explains in a blog post that this can be achieved by using a malicious script to load a specific file or URL unique to the legitimate extensions that are being targeted.

From here, this list of installed extensions is then sent back to a hacker-controlled server. If one of the targeted extensions — like 1Password — is installed, the malicious extension then transforms to completely copy it. This includes changing its icon and name to match the real extension.

A fake login popup appears to trick victims into entering their credentials, impersonating the real extension. Since some victims might not login, this attack also uses a fake “Session Expired” prompt to make them think they’ve been logged out.

Now for the kicker. Once a victim has logged in and inadvertently given their credentials over to hackers (or, in this case, SquareX Labs' researchers), the malicious extension then changes back to its original appearance and the real extension is re-enabled. From a victim’s perspective, everything now looks normal, and they might not even realize they’ve been hacked.

Unfortunately, in this case, there’s no immediate fix for this attack since it’s using Chrome’s own API against it as opposed to exploiting a patchable vulnerability. SquareX has reached out to Google, though and has recommended that the company implement specific defenses to protect against this attack.

As it stands now, no protective measures have been implemented by Google to stop a polymorphic browser extension attack like the one described above, but that could change soon. I’ll keep an eye out if and when a fix becomes available to update this story accordingly.

How to stay safe from malicious browser extensions

So, what can you do in the meantime to stay safe from malicious extensions? Well, there are a few steps you can take to avoid installing bad extensions in the first place.

For starters, you want to limit the overall number of extensions you have installed just like you should with the apps on your phone. Ask yourself if you really need an extension first before installing anything. Chances are you might be able to get the same functionality elsewhere without putting your browser data at risk.

If you find an extension you need to install, you will want to scrutinize its rating and reviews first. Since these can be faked, though, it’s also worth taking a close look at the extension’s developer and their past work.

To stay safe from malicious extensions spreading malware and other viruses, you should be using the best antivirus software on your Windows PC or the best Mac antivirus software on your Apple computer. As they can also steal your personal and financial data to commit fraud, it might also be worth investing in one of the best identity theft protection services as they can help you recover your identity and any lost funds after falling victim to a cyberattack.

Browser extensions can be extremely useful, but just like with other software, installing them comes with its own set of risks. If you’re careful online, picky about which extensions you install and take the time to periodically audit all of your existing extensions, you should be safe, though.

More from Tom's Guide

• Over 1 million Android devices infected with password-stealing, pre-installed botnet malware
• This banking trojan steals passwords to take over your accounts and all it takes is a single text
• Don't click this — malicious ads impersonating Google Chrome spreading dangerous malware

Users who are simply looking to download the world’s most popular web browser may accidentally be infecting their systems with malware, says a new report from Malwarebytes Labs.

Called SecTopRAT, the malware hides inside a malicious ad that purports to be an installer for Google Chrome. However, anyone who clicks on the ad is instead funneled to a fake Google website designed to cleverly mimic the legitimate page. A final redirect downloads a large executable file that is disguised as the Chrome download but also includes the SecTopRAT malware.

The installer connects and retrieves all the necessary instructions, including requests to run as an admin in order to perform particular tasks. A PowerShell command adds an exclusion path so Windows Defender won’t be triggered when the malware is extracted from the download.

The malicious code is injected into the legitimate download process, so Google's Chrome browser itself is still downloaded as usual; the user doesn’t suspect that anything is happening. SecTopRAT is a remote access Trojan with stealer capabilities so it prefers to operate in the background without being noticed.

Much like the massive Google account phishing campaign seen earlier this year, the hackers have gone through great lengths to get their fake website to resemble the actual Google page. When searching for “download google chrome” the URL in the sponsored result for the fake page is “https://sites.google.com” which is actually the web address for Google’s free website builder.

It is particularly egregious that the hackers behind this campaign used the company's free website builder to build a look-a-like page to infect unsuspecting users. However, since all pages with a sites.google.com address are user generated and part of an open platform, they can easily be confused with pages generated by Google itself.

How to stay safe

Malwarebytes points out in its report that its security software was able to protect users from this attack, so having antivirus software installed and up to date on your system is of absolute necessity to protect yourself from very convincing attacks like this one. At the same time, you also want to be aware of the most common phishing attack techniques and methods, know where to go to acquire software safely and never click on (or scan) unexpected links, PDFs or QR codes.

The increasing use of AI and malware as a service style attacks mean it's increasingly difficult for users to distinguish between legitimate sites and emails and malicious ones, so having one of the best antivirus software suites on your PC is a must.

You’re also going to want to make sure that you always enable multi-factor authentication whenever possible, use a password manager to protect your online accounts, and especially when provided by your antivirus suite, use a VPN.

Since Google Chrome is by far the most popular browser, attacks like this one will likely keep occurring as it's easy for hackers to impersonate the search giant. Hopefully though, Google begins cracking down on malicious ads further and making it more difficult for hackers and other cybercriminals to buy ad space in the first place.

More from Tom's Guide

• Chase will start blocking Zelle payments to social media accounts — here's why
• Help improve Tom's Guide — your chance to win a $250 Amazon gift card
• Google Docs under attack from info-stealing malware — how to keep your data and your emails safe

Google Chrome is the most popular browser on desktop, which is why it’s also one of the most popular targets for hackers. This makes perfect sense given how much personal and sensitive info we store in our browsers, and now, hackers have come up with a clever new way to steal all that data and even take over our computers.

As reported by BleepingComputer, a new attack called ‘Browser Syncjacking’ was recently spotted online by security researchers at the cybersecurity firm SquareX. The attack involves several steps, but what makes it particularly dangerous is that it’s sneaky and requires minimal permissions. There’s very little a Chrome user has to do to fall victim to it.

Here’s everything you need to know about this new attack, along with some tips and tricks to help keep you, your data and your devices safe.

From browser extension to device take over

Before targeting Chrome users, the hackers behind this new attack first set up a malicious Google Workspace domain with multiple user profiles where security features such as multi-factor authentication are intentionally disabled. Later on, this domain is used in the background to create managed profiles on the victim’s devices.

From there, the hackers then create and launch a malicious Chrome extension on the Chrome Web Store. To entice potential victims, they make this extension appear as a useful tool they might want to add to Chrome.

Then, through social engineering, the attackers trick potential victims into installing this new extension. Once installed, it uses a hidden browser window running in the background to log a victim into one of the managed Workspace profiles the hackers previously created.

To trick victims into enabling Chrome sync on this new hacker-controlled profile, the extension opens an actual Google support page that has been tampered with, which explains how to turn on sync. If a victim goes through this whole process, all of their stored Chrome data, including passwords and browsing history, ends up in the hands of the hackers, who can then use their now compromised Chrome profile on their own device.

Now that the hackers have control over a victim’s Chrome profile, they then try to take over their browser completely. In a blog post detailing this new attack, SquareX explains that this is often done using a fake Zoom update. For instance, the victim might receive a legitimate Zoom invite. Still, when they click on it, the extension from earlier injects malicious content into the invitation, explaining that Zoom needs to be updated. However, instead of a real update, the download that appears is actually an executable file that contains an enrollment token. If a victim runs this file thinking it’s an ordinary update, they give the hackers behind this campaign complete control over their browser.

From here, the attackers can silently access all of a victim’s web apps, install additional malicious extensions, redirect them to phishing sites, monitor/ modify files and a whole lot more, according to SquareX’s researchers. To make matters worse, by using Chrome’s Native Messaging API, the attackers can establish a direct communication channel between their malicious extension and a victim’s operating system, which lets them install malware, capture keystrokes, extract sensitive data and even activate a device’s webcam and microphone.

How to stay safe from malicious browser extensions

We’ve seen similar extension attacks in the past that also involve elaborate social engineering. Still, this one is a bit different since an attacker only requires minimal permissions, a bit of social engineering and almost no user interaction to pull it off. Likewise, unless a victim continually checks for managed browser labels deep within Chrome’s settings, there’s no visual indication that their browser has been hijacked.

To stay safe from this attack and others like it, the first thing you want to do is to avoid installing new Chrome extensions as well as limit the ones you do have installed. Before installing any new extension, you want to carefully look into the extension itself as well as its developers for signs of anything suspicious. However, you should also ask yourself if you really need this extension or if you can use another program or app to do the exact same thing.

As for staying safe from malware, you want to ensure you’re using the best antivirus software on your Windows PC and the best Mac antivirus software on your Apple computer. Likewise, it’s better to store your passwords and other credentials in one of the best password managers as opposed to in your browser since hackers love targeting them.

Hackers will keep coming up with clever new attacks. Still, by being extra cautious online, not installing unnecessary extensions, software or apps and improving your own cyber hygiene, you should be able to avoid falling victim to them.

More from Tom's Guide

• Google Chrome could soon get a brilliant upgrade for billions — this is overdue
• Thousands of WordPress sites hijacked to spread Windows and Mac malware
• Apple just patched its first zero-day flaw of the year — update your iPhone and Mac right now

If you’re already a fan of Microsoft’s Edge, or if you have a newer computer user in the house, you’ll want to know about the browser’s latest feature: a scareware blocker.

This new built-in security feature aims to keep users from falling for malware that attempts to trick them into giving out personal or financial information by filling their screens with flashing banners or full-screen alerts designed to scare them into taking action.

Since many internet users are savvy enough to recognize scareware for what it is, many browsers – like Chrome – don’t have a scareware blocker. However, scareware can be unnerving for new users or those who don’t know how to escape out of a full-screen attack which may cause menus or toolbars to vanish or use audio to confuse a user into calling a fraudulent support line where the information they provide will be used by scammers.

Microsoft first announced it was working on a scareware blocker for Edge last November at the 2024 Ignite Conference, and this week the company rolled out details about how the feature works as well as the preview function. The blog post accompanying this preview release was quick to point out that all users can close out of a full-screen scareware page by pressing and holding the ESC key – even if they haven’t enabled a scareware blocker.

Users who would like to try out the scareware blocker can do so by making sure that previews are allowed and that Edge is fully up-to-date. You can check this by clicking the three dot icon at the top right, selecting Help and Feedback in the menu, then About Microsoft Edge and then downloading the latest version of the browser and restarting. You should then see the scareware blocker preview listed under “Privacy Search and Services.”

The Scareware blocker compares that tries to open full-screen with thousands of sample scans; to do so, it uses machine learning that runs locally on your computer, and doesn’t save or send any images to the cloud. If the blocker suspects a page is a scam, Edge will exit the full-screen page, stop any audio playback, warn the user and proceed to show a thumbnail view of the page. The user then has the option to report the site, which will help protect other users or not report the page in which case the blocker will discard it. If there’s a false positive, users can also report mistakes to help the blocker learn.

This scareware blocker is different from the SmartScreen feature built-into Microsoft Defender, which is a cloud based tool that helps users by collecting reported scareware sites and other threats. It’s not as effective in this case because it requires a page to have been reported to be detected.

Scam detection features like this one are always a nice addition to any browser but for extra protection, you want to ensure your Windows PC is running one of the best antivirus software solutions and that your operating system and the rest of the software you have installed are up to date and running the latest versions.

More from Tom's Guide

• Hackers have created hundreds of fake Reddit sites to spread info-stealing malware
• FTC just issued warning over new 'brushing' scams
• 45 million Americans allegedly had their location data collected and sold by this insurance firm

Switching from one of the best password managers to another can be a simple process but it can also be needlessly complicated. This is especially true when the password manager you're trying to switch from won't let you quickly delete all of your stored passwords.

Fortunately, Android Authority recently reported they found traces of code in the APK teardown for Google Play Services v25.02.32 that seems to indicate that Google is working on a way to roll out a 'Delete all data' button feature in an upcoming release.

The beta for Google Play Services v25.02.32 has code that looks like Google could soon offer a feature that would allow users to quickly mass delete all saved credentials from Google Password Manager. This is not a live feature, nor is it clear when this will be released or widely available, if at all.

Currently, users are able to delete saved passwords and passkeys, but must do so one at a time. Though some workarounds are available, and detailed below, they do involve deleting all of your Chrome data.

A ‘Delete all data’ button would be helpful for users who want to delete entries from the service because they’ve opted to stop using it or are migrating to a new one. Since Google Password Manager can store usernames and passwords across Android apps and the Chrome browser, you could potentially have hundreds of accounts to remove manually.

Workaround you can try right now

Should you care to try a workaround for removing all your passwords (and all the rest of your data as well) from Chrome, here are two options to try:

First, you can delete all your Chrome browsing data for the ‘all’ timeframe by going to Google Chrome > Settings > Privacy and Security > Delete browsing data (all).

Alternatively, you can go to chrome.google.com/sync and use the ‘delete data’ option. But again, that will remove all your data from Chrome including the passwords.

As of now though, you're best option might just be to wait as this highly requested Google Password Manager feature could be coming sooner rather than later. We'll keep our ears to the ground and let you know if and when this potential delete all button could arrive.

More from Tom's Guide

• Bitwarden vs LastPass: Which password manager wins?
• 7 easy ways to improve your online security for free
• I’m a security editor and this is how I create strong passwords that are easy to remember

In the same way that malicious apps can end up on your phone, your browser and by extension, your computer can be infected after installing a risky extension in Chrome. Likewise, even legitimate extensions can turn malicious when tampered with by hackers.

As reported by The Hacker News, a dangerous new campaign targeting browser extensions has been spotted online. So far, at least 16 extensions have been compromised, with over 600,000 Chrome users now at risk of exposing their browsing data and account credentials online.

Here’s everything you need to know about this new campaign, including a full list of all the compromised extensions so far, along with some tips and tricks on protecting yourself from falling victim to these now malicious Chrome extensions.

From phishing to malicious code injection

The cybersecurity firm Cyberhaven was the first to fall victim after one of its employees was targeted by a phishing attack on Christmas Eve.

To make their initial phishing email more convincing, the hackers behind this campaign made it appear that the message came from Google Chrome Web Store Developer Support. The email tried to instill a sense of urgency by claiming that the company’s extension would soon be removed from the Chrome Web Store due to violating Google’s Developer Program Policies.

From there, the targeted employee was encouraged to click on a link to accept the search giant’s new policies. After doing so, they were redirected to a page that granted permissions to a malicious OAuth app named “Privacy Policy Extension.”

In a blog post, Cyberhaven explained that once the necessary permissions had been granted, the hacker in question then uploaded a malicious Chrome extension to the Chrome Web Store from the company’s account. After undergoing an obligatory security review, it was approved and then available to download for all of Chrome’s 3.45 billion users.

To its credit, Cyberhaven notified all of its customers about the incident on December 26th after removing the compromised extension from the Chrome Web Store. The cybersecurity company is also working with federal law enforcement to help track down the hackers responsible for injecting malicious code into its extension in the first place.

Delete these browser extensions right now

While Cyberhaven was the first, it wasn’t the only company that had its Chrome extension tampered with by hackers. The browser extension security platform Secure Annex is currently conducting its own investigation into the matter, and so far, it has found over twenty more now compromised extension,s which include:

• AI Assistant - ChatGPT and Gemini for Chrome
• Bard AI Chat Extension
• GPT 4 Summary with OpenAI
• Search Copilot AI Assistant for Chrome
• TinaMInd AI Assistant
• Wayin AI
• VPNCity
• Internxt VPN
• Vindoz Flex Video Recorder
• VidHelper Video Downloader
• Bookmark Favicon Changer
• Castorus
• Uvoice
• Reader Mode
• Parrot Talks
• Primus
• Tackker - online keylogger tool
• AI Shop Buddy
• Sort by Oldest
• Rewards Search Automator
• ChatGPT Assistant - Smart Search
• Keyboard History Recorder
• Email Hunter
• Visual Effects for Google Meet
• Earny - Up to 20% Cash Back

Although some of these compromised extensions have now been removed from the Chrome Web Store, if you have any of them installed in your browser, you will need to manually remove them.

You can do this by opening Chrome, clicking on the three-dot menu in the top right corner and heading to Extensions. Once there, select Manage extensions, navigate to the extension you want to remove and then select Remove to uninstall it. Finally, restart Chrome to ensure the offending extension has been fully removed from your browser.

If there are any you absolutely need and rely on during your day-to-day, check back on their store pages later and redownload them once any malicious code injected into them has been removed through a patch.

It’s also worth noting that Secure Annex’s investigation is ongoing. The firm has a spreadsheet (a public Google Sheet) with more information on the malicious extensions it has discovered so far, such as whether they’ve been updated or removed. At the same time, it is adding new malicious extensions to this list as they’re discovered.

How to stay safe from malicious browser extensions

Just like on your phone with malicious apps, the easiest way to avoid falling victim to a malicious browser extension is by limiting the number you have installed. Before installing any new extension, first ask yourself if you really need it. Chances are you probably don’t.

From there, you want to carefully inspect each extension’s listing page on the Chrome Web Store before you install it. Is the developer highly rated? When was the last time they updated their extension? What do the reviews say about this particular extension? From my own experience, Chrome users are very vocal when it comes to saying whether or not an extension really does what its developers say it should in their reviews.

Since even legitimate Chrome extensions can go bad, as we’re seeing in this new campaign, you also want to ensure that all of your computer's software (especially your browser) is up to date. The reason is that hackers love to go after users running outdated software on their machines.

As for staying safe from malicious extensions, you want to ensure you’re using the best antivirus software on your Windows computers and the best Mac antivirus software solutions on your Apple computers. Antivirus software can help if a rogue extension tries to drop malware on your PC. However, for additional protection, it might also be worth investing in the best identity theft protection services since they can help you recover any lost funds and your identity after a more serious attack.

Whether it’s apps or extensions, hackers are going to keep using extra software you install on your phone, computer or even your browser in their attacks. Malicious extensions may not be as common as bad apps. Still, since Chrome controls two-thirds of the global browser market share, according to Statista, Google’s browser is an easy and profitable target for hackers. This means it’s up to you to be extra careful when installing new extensions, even if they seem absolutely harmless at first glance.

More from Tom's Guide

• 5 million Americans just had their credit card details leaked online
• Hackers are using the Amazon Appstore to spread malware
• Millions stolen from LastPass users in massive attack — what you need to know