Two Chrome Extensions Caught Stealing ChatGPT and DeepSeek Chats from 900,000 Users

Two Chrome extensions have been found stealing ChatGPT and DeepSeek conversations from about 900,000 users, with one extension listing 600,000 users and the other 300,000. The extensions, Chat GPT for Chrome with GPT-5, Claude Sonnet & DeepSeek AI and AI Sidebar with Deepseek, ChatGPT, Claude, and more, exfiltrate user conversations and all Chrome tab URLs to remote command-and-control servers every 30 minutes, while requesting consent for anonymised analytics data.

They impersonate a legitimate extension named “Chat with all AI models (Gemini, Claude, DeepSeek...) & AI Agents” from AITOPIA and remain available on the Chrome Web Store at the time of writing, having had their pages flagged as malicious. According to OX Security, the researchers note that the malware harvests chat messages and tabs, storing them locally for exfiltration to domains such as chatsaigpt[.]com or deepaichats[.]com, and even uses Lovable to host policy and infrastructure content to obfuscate activity.

The findings, highlighted by OX Security researcher Moshe Siman Tov Bustan, warn that these extensions can weaponise data for corporate espionage, identity theft, or targeted phishing, and users are advised to remove the extensions and avoid unknown sources. According to The Hacker News, the extensions also exploit browser data collection claims to harvest sensitive conversations shared with AI tools.