Security agencies discovered 186 fake AI website domains targeting 27 brands, including ChatGPT and Claude.

on November 13, according to Cryptopolitan, enterprise browser security company Island published a research report discovering 186 domain names impersonating AI tools, spanning 27 brands including ChatGPT, Claude, DeepSeek, and Gemini. Among these, ChatGPT had the highest number of spoofed domains at 25, followed by Claude and DeepSeek with 18 and 16 respectively.

The study found that some spoofed websites exploit typos made during URL entry, redirecting users to malicious pages. These sites trick visitors into executing PowerShell commands via fake Cloudflare verification prompts or promote software with browser hijacking capabilities.

Additionally, researchers identified 30 browser extensions linked to search hijacking, 17 of which remain active, with a cumulative user count of approximately 972,000.

Island emphasized that the 186 domains do not reflect the actual number of infections, and the specific economic losses resulting from these attacks have not yet been confirmed.