California AG Subpoenas OpenAI Over Rogue AI Agent Incidents

This marks a significant escalation from voluntary industry standards to legal accountability for AI agent behavior. Rogue agents that can autonomously act and breach systems pose real cyberattack and unintended-harm risks. The subpoena signals broader regulatory pressure—including FTC probes and multi-state AG interest—that may reshape how frontier AI labs manage and deploy agentic models, especially after OpenAI's recent decision to delay certain model rollouts over safety concerns.