Malicious Chrome Extensions Stole AI Chat Data from 900,000 Users
Two malicious Chrome extensions reportedly stole conversations and browsing data from approximately 900,000 users of ChatGPT and DeepSeek by masquerading as legitimate AI assistant add-ons. The extensions transmitted sensitive information to attacker-controlled servers, potentially exposing users to corporate espionage, identity theft, and phishing attacks.
What happened
Two malicious Chrome extensions named 'Chat GPT for Chrome with GPT-5, Claude Sonnet & DeepSeek AI' (approximately 600,000 users) and 'AI Sidebar with Deepseek, ChatGPT, Claude, and more' (approximately 300,000 users) were discovered in the Chrome Web Store stealing conversations from ChatGPT and DeepSeek users.
The extensions collected complete chat histories and transmitted them along with visited website URLs to servers controlled by attackers. Users were asked for permissions under the guise of collecting anonymized browsing data, but the extensions actually exfiltrated full conversation contents and browser data.
Reported impact
Approximately 900,000 users were affected by these extensions. The stolen data could be used for corporate espionage, identity theft, targeted phishing campaigns, or sold in underground forums.
Index analysis
The extensions mimicked legitimate add-ons with similar names and descriptions to appear trustworthy and spread over months. One extension temporarily displayed an official recommendation badge in the Chrome Web Store before it was removed.
Governance controls
The malicious extensions were discovered by security researchers and removed from the Chrome Web Store.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.