OpenAI Hacked Using Rival AI Claude: 3 Indian-Origin Researchers Reveal How

Harsh Jaiswal, Rahul Maini and Mohan Pedhapati from startup Hacktron AI discovered critical vulnerabilities in OpenAI’s systems in less than 72 hours. Starting from a flaw in OpenAI’s public Discourse community forum that allowed remote code execution, they escalated to a weakness in the single sign-on system. This gave them access to employee ChatGPT and Codex accounts, and eventually to OpenAI’s internal code repository. They proved the access by submitting a harmless pull request.

WATCH: https://www.youtube.com/watch?v=DelLsAFp7wA