Fake ChatGPT Ads on Google Push Malware Through Sponsored Links
A paid search result posing as ChatGPT is installing malware on users who click it, according to a new report on the scam.
A malware campaign is buying Google ads that impersonate ChatGPT, and the people clicking through are not getting the chatbot they expect. The scheme surfaced this week in a ZDNet report, which traced the trap back to a sponsored link sitting at the top of search results. Because the ad carries Google's paid placement label, many users treat it as a trusted route to OpenAI's product rather than an unknown third party.
The mechanics are straightforward. A victim searches for ChatGPT, clicks the top sponsored result, and lands on a page built to look like the real service. Instead of a chat window, the site pushes a download, which installs malicious software on the machine. The report describes the fake page as convincing enough that the mismatch only becomes clear after the fact, which is what makes the ad format so effective as a delivery vehicle.
Search ads have become a favored channel for this kind of attack because they borrow credibility from the platform serving them. Google labels sponsored results, but the label says paid placement, not vetted destination, and scammers have repeatedly exploited that gap around high-demand software. ChatGPT's popularity makes it an obvious target: huge search volume, a mostly non-technical audience, and a product that people expect to sign up for quickly.
For defenders, the practical advice is to skip the ad slot entirely. Typing the address directly, or using a bookmark saved from a previous legitimate visit, removes the intermediary that attackers depend on. Users should also be suspicious of any ChatGPT page that asks for a file download, since the service runs in a browser and does not require installing an executable to start a conversation.
The story fits a broader pattern of AI-branded fraud rather than an isolated incident. Scammers have spent the past two years attaching OpenAI's name to fake apps, browser extensions and crypto schemes, betting that familiarity with the brand will lower a victim's guard. Google and OpenAI have not detailed what, if anything, they are doing about this specific ad cluster, and the report does not say how many people were affected or where the downloaded payload leads.
What to watch next is whether Google tightens verification for ads that mimic well-known software brands, and whether OpenAI pushes search engines harder to police impersonation of its name. Until either changes, the sponsored slot remains the cheapest way for an attacker to stand in front of a user who is already looking for ChatGPT.