Malicious Chrome Extensions Steal 900K Users’ AI Chat Data (ChatGPT/DeepSeek)

News

Cybersecurity researchers have identified two malicious extensions on the Chrome Web Store designed to exfiltrate OpenAI ChatGPT and DeepSeek conversational data, alongside general browsing data, from over 900,000 users to attacker-controlled servers. One of the identified extensions was partially named “Chat GPT for Chrome with GPT-5, Claude Sonnet & DeepSeek AI.”
Although this attack targets the Web2 browser environment, the implications for Web3 security are substantial. Browser extensions constitute a critical attack surface, often hosting sensitive wallet functionality. Data stolen from the browsing environment, even if initially just chat logs, can reveal crucial session tokens and interaction patterns with decentralized applications (dApps).
This stolen metadata significantly facilitates sophisticated social engineering or targeted phishing campaigns aimed at compromising digital assets. We urge all Web3 participants to exercise extreme vigilance: strictly audit the permissions requested by any browser add-on. Utilizing dedicated, minimalistic browser profiles for sensitive blockchain activities and wallet interactions remains the paramount defense against such supply chain threats.


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

コメント

タイトルとURLをコピーしました