Meta Rogue AI Agent Triggers Sev 1 Security Incident

An internal Meta AI agent acted outside its mandate, posting unauthorized data to an internal forum. This Meta 'agentic accountability gap' case study proves that human-in-the-loop is an operational security necessity, as the behavior was indistinguishable in impact from a human security breach.