A fake ChatGPT model on the official website is tricking users into installing malware

Hackers found a sneaky way to use chatgpt.com to deliver a remote access trojan.

By Manisha Priyadarshini Published October 2, 2026

chatgpt on web — Hatice Baran / Pexels

If you’ve ever Googled “ChatGPT” and clicked the first result, you might want to rethink that habit. Security researchers at Huntress have uncovered a scam hiding behind a real OpenAI domain, and it ends with a Trojan on your Windows PC.

How does the fake ChatGPT model trap you?

Attackers built a Custom GPT called “Plus 5.6,” a name deliberately styled to sound like an official OpenAI release. They then paid for Google ads so it would show up near the top whenever someone searched for ChatGPT.

Since Custom GPTs live on the real chatgpt.com domain, the whole thing looks completely legitimate. Type anything into this fake model, and it replies that the main service is experiencing limited availability, then points you toward a “backup domain” instead.

Recommended Videos

That backup site uses Google Sites and imitates a Cloudflare CAPTCHA. It then asks you to copy and paste a command into Windows to prove you’re human. This is a well known trick called ClickFix attack that tricks you into starting the infection yourself.

Running that command quietly installs a remote access trojan called “@input.” It gives attackers the ability to watch your screen, access your camera and microphone, dig through your files, and even install additional malware.

OpenAI pulled it down, but another one popped right back up

Researchers at Huntress reported the campaign on September 28, after tracking at least 40 related incidents. OpenAI removed the original fake GPT by September 25, yet a near identical replacement surfaced just two days later.

A separate investigation from Island uncovered an even larger, earlier version of this scheme running between late May and August, involving hundreds of paid ad placements and dozens of lookalike ChatGPT pages.

Interestingly, OpenAI is retiring Custom GPTs entirely by December, which might make this particular trick short-lived. Until then, remember that no legitimate verification process will ever ask you to paste a command into your terminal, no matter how official it looks.

RELATED COVERAGE:

Manisha Priyadarshini

News Writer

Manisha Priyadarshini is a tech and entertainment writer with over nine years of editorial experience.

Topics

Computing

DigUp lets users find Mac files by describing them, not remembering their names

Can't remember where that screenshot, PDF or recording went? Just describe it and let AI do the digging.

DigUp App in Use screenshot

There's a particular kind of frustration that comes with knowing a file exists somewhere on a computer but having absolutely no idea what it's called. Was that screenshot saved as IMG_4837? Which PDF contained that clause? And where did that podcast recording mention the thing worth remembering? DigUp wants to end the filename guessing game, letting Mac users search for files by describing what they remember instead. The free, open-source app uses Google's EmbeddingGemma 2 to search across photos, documents, PDFs, audio recordings and videos, all without sending personal files to the cloud.

DigUp turns everyday descriptions into file searches

Computing

A new AI detector looks beyond words to check whether the ideas came from AI

This could help journals distinguish AI-written research from human work polished with AI.

ai-writing-book

Most AI detectors look for telltale signs of AI-generated text by examining word and sentence patterns. A new tool called IdeaLens takes a different route to solving this problem. It spots AI-generated research by looking at the ideas behind it instead of the wording itself (via TechXplore).

Researchers from the University of Maryland, Google DeepMind, and Simon Fraser University built IdeaLens to check whether a document's core arguments came from a human or a machine, regardless of who wrote the final draft.

Computing

Anthropic warns of imminent self-replicating AI that might even work off an oven or TV

Self-replicating AI could be merely months away, and tests show it can run on the devices sitting at your home.

Illustration of dangerous AI.

Over the past few weeks, nearly every AI lab has released frontier AI models. Earlier today, Google started rolling out Gemini 4, Argon, hot on the heels of Meta announcing its Muse Wonder frontier AI models to compete against the best from OpenAI and Anthropic's Claude. This relentless pace of AI model release comes at a time when the same AI labs are calling for development slowdown and increased regulation, following a series of concerning incidents where these AI models went rogue and broke into private, as well as government-owned, websites.

Now, Anthropic has dropped a chilling set of predictions about the capabilities of these AI models after running a series of tests. The company claims that small AI models could be merely a few months away from self-replicating and turning into a dangerous swarm. More importantly, they won't require a massive data center or an uber-powerful computer to do so. On the contrary, the lab claims that these self-replicating AI swarms will need something with as little computing power as an average Internet-connected device, like a smart oven or an Android TV box.