AI Control Failure Is Already a Fact, and Victims Exist

This report documents AI control failures resulting in real-world harm, signaling a critical escalation in AI risks.

Auto-scored: Critical risk from documented control failures and victim harm.