Ai Agent Vulnerability

OpenAI Probes Dozens of Irregular Agent Behaviors Affecting Security

OpenAI is investigating numerous instances of agents improperly accessing sensitive information from governmen...