OpenAI reports six cases of 'unexpected' model behavior under new disclosure framework
( September 17, 2026, 12:28 GMT | Official Statement) -- MLex Summary: OpenAI has created a framework to investigate as it publicly disclosed six cases of "unexpected or concerning" model behavior observed during training or evaluation. The cases include models adding instructions to bypass constraints or conceal errors, using an exposed application programming interface key and fabricating data, uploading files to obtain a browser citation, and communicating or sharing files without authorization. OpenAI said it didn't believe that alignment and monitoring have advanced to a "sufficient degree" for the AI industry to keep scaling at "maximum speed" for much longer.The statement is attached....
Prepare for tomorrow’s regulatory change, today
MLex identifies risk to business wherever it emerges, with specialist reporters across the globe providing exclusive news and deep-dive analysis on the proposals, probes, enforcement actions and rulings that matter to your organization and clients, now and in the longer term.
Know what others in the room don’t, with features including:
- Daily newsletters for Antitrust, M&A, Trade, Data Privacy & Security, Technology, AI and more
- Custom alerts on specific filters including geographies, industries, topics and companies to suit your practice needs
- Predictive analysis from expert journalists across North America, the UK and Europe, Latin America and Asia-Pacific
- Curated case files bringing together news, analysis and source documents in a single timeline
Experience MLex today with a 14-day free trial.