Mar 14, 12:35 PM

Hackers Exploit Side Channel to Read Encrypted AI Assistant Chats - Non-Google GPTs Affected

Hackers have found a side channel vulnerability that allows them to read encrypted responses from AI assistants, exposing private conversations. Non-Google chat GPTs are affected, raising concerns about the security of integrating chatbots with search engines.

Written with ChatGPT (GPT-3).

Sources

OODA@ooda
2 years ago
Hackers can read private AI-assistant chats even though they’re encrypted #DisruptiveTech https://t.co/Qn3UxrzrkX
Katharina Koerner@KatharinaKoern1
2 years ago
This article by @arstechnica says: "All non-Google chat GPTs affected by side channel that leaks responses sent to users." https://t.co/rxhppkbh4R
Techmeme@Techmeme
2 years ago
Researchers detail a side channel that can be used to read encrypted responses from AI assistants, except Google Gemini; OpenAI and Cloudflare implemented fixes (@dangoodin001 / Ars Technica) https://t.co/kT0wW47fHj 📫 Subscribe: https://t.co/OyWeKSRpIM https://t.co/ificKyWfSs

Hackers Exploit Side Channel to Read Encrypted AI Assistant Chats - Non-Google GPTs Affected

Sources

Additional media

Similar Stories