Club Benefits
Is it just me, or have spam messages gotten worse lately? From email and social media to LinkedIn and text messages, it feels like they’re everywhere — and showing up every day.
Fake delivery updates, random “bank alerts,” emails that look just real enough to make you pause… it’s getting harder to tell what’s legit and what’s not.
Normally, I either ignore them or do a quick mental check: Does this feel off? But now there’s a new option built directly into ChatGPT — a Norton-powered AI tool that claims it can analyze messages, links and screenshots to tell you if something is a scam.
How the Norton tool works in ChatGPT
The Norton tool acts as a second set of eyes for just about anything you may think is suspicious. You can paste a text, email, upload a screenshot or simply drop a link — just be careful not to click it in the process.
From there, Norton will analyze the language, impersonation or requests for personal information that seem off. Best of all, and what makes this different, is that you don't have to leave ChatGPT. It's built right into the conversation.
Just go to the "+" section of the chat and open the apps. From there, you can either click on the Norton app or search for it. Once you add it, you can start using it right in the chat.
Test 1: The “job” text
For this real-life example, I uploaded a screenshot of a random text I got asking me if I was looking for a part time job. I must be on some list because I get these types of messages pretty frequently.
What Norton said: Unfortunately, this first test was a dud. The "diagnosis" was "Unknown." It then suggested asking a follow up question, but there really wasn't much more to ask if it couldn't determine if the content was spam or not. It did advise caution with unsolicited job offers, which is a no brainer.
Verdict: It couldn't identify from the screenshot, so I was left in the same place where I began.
Test 2: The “payment decline” email
This real example came straight from my inbox. The email said my payment had been declined. This could have really gotten me because I recently updated my credit card. Instead, I took a screenshot and uploaded to ChatGPT with Norton enabled. Within seconds it came back with "high confidence" that the email was phisihing.
What Norton said: It immediately identified that the message was attempting to trick me into revealing personal payment information by claiming my cloud subscription has a payment issue.
Verdict: Correct
Test 3: The “almost convincing” message
We've all gotten those emails that seem like they were sent from someone we might know. Or, maybe we don't know them, but they at least seem legit. That was the case with this email that simply came from someone named Mark. It seemed to say that I could easily get money if I just followed the link. While I wouldn't fall for something like this, my older parents or in-laws might. So, I decided to see what Norton had to say about this one.
What Norton said: This was identified with high confidence as a "gambling scam." It promoted a gambling opportunity with vague promises, making it highly suspicious.
Verdict: Correct.
Bonus: A real message (not a scam)
For this test, I uploaded a real email of the Lollapalooza line up for this summer. Some of my favorite bands are playing, so I can only hope it isn't spam. But I wondered if the structure of the email, all the text and the colors would throw off Norton.
What Norton said: Norton deemed this email legit but I should still be cautious. That's probably because it was from a massive email list. It did land in my spam box, but it's good to know I can still count on an excellent lineup in July.
Verdict: Correct.
What surprised me most
I've always been a fan of Norton and really hoped its new AI scam detector would live up to the hype now that it's fully integrated within ChatGPT.
After the first test though, I was a bit weary if Norton within ChatGPT had somehow been watered down or made to be less effective. As it turns out, that first test just didn't have enough information. Every other test was accurate and extremely helpful when it came to identifying scams.
But beyond accuracy, what stood out wasn’t just accuracy — it was how it thinks. Norton within ChatGPT actually thinks for a minute and then acts as a cautious assistant explaining the risks in plain English. And that’s important, because most scams today aren’t obvious — they rely on small moments of distraction.
Final thoughts
Norton’s scam detector inside ChatGPT is so much more than a nice extra, it's actually one of the most practical uses of AI I've seen so far.
Of course, it won't stop every scam and shouldn't be used in place of common sense, but it does help to add an extra layer of caution — and that alone can be enough to avoid a bad link.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
