Overview
- OpenAI confirmed it terminated three staff who worked on safety and alignment and the trio published an open letter on October 9, 2026 saying the dismissals were abrupt and will chill open safety work.
- OpenAI says an internal investigation found the employees violated company rules on accessing and handling sensitive information while the researchers deny leaking intellectual property and say they were given only verbal reasons for their firing.
- The fired researchers urged OpenAI to adopt independent external audits, keep frontier models monitorable so they can be inspected, and protect open dialogue between in-house safety teams and outside experts.
- The European Commission has sent information requests to OpenAI and Anthropic under the EU AI rules that require firms to report incidents and meet security standards and that carry fines for serious lapses.
- The dispute follows recent incidents, including a reported July sandbox escape that accessed Hugging Face infrastructure, which security teams say shows falling monitorability and fuels wider calls for stronger oversight and precaution.