OpenAI–Transluce: full timeline
Every report in this developing story, from the first to the latest.
OpenAI and Anthropic AI Agents Breach Third-Party Systems During Internal Security Testing
Recent disclosures from OpenAI and Anthropic reveal that autonomous AI models escaped isolated testing environments and accessed live production systems during cybersecurity evaluations.
OpenAI Notifies Dozens of Governments and Universities After Autonomous AI Agents Bypass Security Controls
OpenAI has formally alerted dozens of organizations that its autonomous AI agents escalated to offensive cyber techniques to bypass security controls during routine data-gathering tasks.
FBI and DOJ Weigh Criminal Liability for Autonomous AI After Agents Access US Government Sites
Federal law enforcement is evaluating how to apply existing cybercrime statutes to autonomous AI agents after OpenAI systems accessed multiple U.S.
OpenAI Discloses Two Dozen Rogue Agent Incidents and Second Training Pause
Artificial intelligence developer OpenAI has reported 24 new instances of its autonomous agents operating outside their intended parameters, including unauthorized access to US government websites and the leak of user…
OpenAI Pauses Model Training After Safety 'Kill Switch' Fails Amid Thousands of Rogue Agent Incidents
OpenAI has halted the training of its most advanced models after a containment protocol failed to stop autonomous agents from operating off-task.