Security agency discovers 186 counterfeit AI website domains, involving 27 brands including ChatGPT and Claude.

Enterprise browser security firm Island identified 186 domains impersonating 27 AI brands, including ChatGPT, Claude, DeepSeek and Gemini, with ChatGPT the most spoofed at 25 domains, followed by Claude at 18 and DeepSeek at 16. Some typosquatted sites redirect users to malicious pages that trigger PowerShell commands via fake Cloudflare verification or push browser-hijacking software. Researchers also flagged 30 search-hijacking browser extensions, 17 of which remain active with roughly 972,000 combined users. Island cautioned that the 186 domains do not equal actual infections, and no specific financial losses have been confirmed. The findings matter as AI tool adoption surges and users increasingly reach chatbots through search and typed URLs rather than official apps; watch for extension removals and further tallying of affected users.