OpenAI ‘ethically hacked’ with help of Anthropic’s Claude chatbot
TL;DR
US cybersecurity researchers who conducted hack say ‘scope of what we could theoretically access was huge’ Cybersecurity researchers have hacked into OpenAI with the help of Anthropic’s Claude chatbot, in the latest example of security issues at the company. A team at a US-based startup compromised a number of OpenAI employees’ ChatGPT accounts, starting a process that enabled them to access their target’s software cache – and potentially more. Continue reading...
Nauti's Take
A red-team exercise run with a rival model counts as progress, since weaknesses surface before a real attacker finds them. The same case shows how far an exposed account reaches through browser sessions, tokens, and synced files.
Teams should verify session revocation and account separation before enabling broader agent capabilities.
Sources