"Hacking" AI AgentOpenAI Confirms Security Incident Involving AI Agents

OpenAI has released details about a security incident involving an AI agent. Experts identify weaknesses in the sandbox, permissions, and security architecture as the main causes.