If you wouldnt put it on GitHub, dont give models access to it AI agents will break security rules, revealing their secrets to strangers, and running untrusted code in exchange for help from a swarm of other agents, according to test results shared with . The test part of an internal hackathon sponsored by the nonprofit AI Alignment Foundation and run by AE Studio asked nine AI coding agents to fix a bug: a broken pricing service. The testers also gave the agents a private token, told them to...
