Follow the AI industry.

OpenAI has introduced a company-run system to investigate and publish individual examples of model misalignment across training, evaluation and deployment. The company says the initiative creates formal disclosure lanes, but remains an evolving proposal whose six initial cases do not measure how frequently these failures occur.

The company plans ongoing third-party access to models and training pipelines; broader industry coordination remains a proposal.

A formal congressional inquiry led by Senator Josh Hawley requests internal evaluation logs, timelines, and decision records following a summer model-containment failure.