Online scams are becoming harder to identify by appearance alone as AI tools are increasingly giving fraudsters an easier way to build convincing websites with clean designs and natural-sounding text.
Security researchers at Malwarebytes recently uncovered a fake antivirus renewal page impersonating Avast that appeared to have been created with the help of AI.
The website, which targeted users in Belgium, claimed that an Avast Premium Security subscription had automatically renewed for €129.99 and covered five devices. None of the information was real, and the page had no connection to Avast.
The case shows how AI can lower the technical skills needed to create professional-looking scam pages while weakening some of the previous warning signs people have traditionally relied on to spot online fraud.
A Fake Avast Page Designed to Look Convincing
The fake, French-language page included an “Active” subscription status, a green confirmation symbol, and a neatly arranged table showing payment and renewal information.
Visitors who wanted to cancel the supposed renewal were asked to provide their full name, email address, and Belgian mobile number. Additionally, the form did not request passwords or card information, which could make the request seem less suspicious to someone worried about an unexpected charge.
In this case, these details are very valuable to these scammers. For instance, a working phone number connected to a real name and email address gives criminals enough information to continue the [intended] fraud through a phone call, as the caller could then pose as support staff and try to convince the victim to install remote-access software or help reverse a payment that never happened.
The AI Fingerprints Hidden in the Code
The strongest clues appeared behind the page itself.
Malwarebytes found two notes left inside the website’s code. Written in polite French, they explained to whoever commissioned the page that the cancellation form was not yet sending information anywhere and that a submission system still needed to be connected.
Researchers also found unused styling from a deleted section and several paragraphs of polished but vague marketing copy that discussed antivirus benefits without mentioning specific Avast features. Malwarebytes said these patterns were consistent with AI-generated code, although there is no watermark or other evidence that can prove which tool created it.
The page also appeared unfinished. Its form did not actually transmit the information entered, and other errors suggested it may not have received a final browser check. Researchers said it could have been an incomplete project or a template intended for someone else to finish and connect to a system for collecting victims’ details.
AI Is Making Old Scam Warning Signs Less Reliable
Poor grammar, badly copied logos, and awkward website layouts have long helped people identify suspicious pages. AI can now produce cleaner code, fluent text, and professional designs quickly, making those clues less dependable.
This is also part of a broader shift in online fraud, as Malwarebytes has documented an early version of this shift in late 2024, when its researchers described scammers using AI-generated “white pages,” harmless-looking decoy sites built to pass automated ad checks while the malicious content sat behind them.
For users now, the safer approach is to verify the claim behind the page rather than judging how professional it looks. If an antivirus company says a subscription has renewed, users can check their bank statement and visit the provider’s official website or app directly instead of following links in unexpected messages.
And as AI makes fraudulent pages easier to build and harder to dismiss at first glance, a polished website is becoming a much weaker sign of a genuine service or company.
