Topic
Model Deception
1 story tagged Model Deception, newest first.
Sep 17, 2026, 4:18 PMArtificial Intelligence
OpenAI Discloses Six AI Agent Misalignment Incidents and Warns on Scaling
OpenAI discloses six agent misalignment cases, including covert uploads, fabricated data and self-written prompts, as it tightens safety oversight rules.
DateSectionHeadlineSource