Fake AI Apps and Malicious Browser Extensions: How to Spot Them Before You Install

What a fake AI app or extension does

AI assistants are popular, so criminals and careless developers copy their names. A listing called “ChatGPT for Chrome” or an app with a familiar logo may have no link to the company it imitates. What it does next depends on who built it.

Malwarebytes, in a November 2025 review of app store clones, sorted them into three groups. Some are harmless wrappers that pass your questions to a real AI service. Some are adware impersonators that borrow a brand’s look to show ads and send data to advertisers. The third group is malware dressed up as an AI tool.

The extension case

Browser extensions are the clearest example. In January 2026, SecurityWeek reported research from OX Security on two Chrome extensions that impersonated a real extension called AITOPIA. They had more than 900,000 combined downloads. According to that report, a consent prompt asked to collect “anonymous, non-identifiable analytics data”, while the extensions collected full ChatGPT and DeepSeek conversations along with URLs from open tabs and search queries.

In March 2026, Microsoft Defender published its own analysis. It described malicious Chromium extensions posing as AI assistants and said it saw activity in more than 20,000 enterprise tenants. Microsoft reported that they copied the branding and permission prompts of a legitimate extension and were distributed through the Chrome Web Store, and that Edge users could install them too. It also reported that data collection came back on after updates even for users who had declined it. Microsoft’s post says reporting cited about 900,000 installs.

Why does this matter to you? A chat history can hold contract drafts, medical questions, passwords pasted in a hurry and customer data. URLs can contain session tokens. Stolen information like that fuels convincing phishing messages that mention details only you and the AI knew.

How to check before you install

• Start at the vendor’s website. Use the download link there, not an ad, a social media post or a search result.
• Check the developer name. Malwarebytes advises looking for small changes such as extra letters or punctuation.
• Read reviews for repeated complaints about pop-ups, ads or unexpected charges.
• Read the permissions. A sidebar chatbot does not need to see every page you visit, although some extensions do ask for that. Ask yourself whether the benefit is worth it.
• Do not trust a store listing alone. Both extensions in the OX Security research were available in the Chrome Web Store before they were removed.

If you already installed one

Remove it from your browser’s extensions page, then change the passwords on accounts you used while it was active, starting with email and anything that holds money. Turn on two-factor authentication where you can. Watch for unusual messages that reference your private chats. If it happened on a work device, tell your IT team, since Microsoft’s advice to organisations is to inventory and audit extensions.

Our data breach checklist covers the next steps if sensitive material was in those chats. For habits that limit the damage in any case, see what not to tell AI chatbots, and read how AI agent permissions widen the risk when a tool can act on your behalf.

Fake apps are not new, but AI tools give them a better disguise, because people expect to type private things into an assistant. Treat every new AI install as a decision about who gets to read your conversations.

This article reflects public information as of 2026-11-06. AI tools and rules change quickly; we update articles when new facts are published. Spot an error? See our editorial and corrections policy.