OpenAI fires safety researcher who says he was sacked for talking to auditors

Tomek Korbak: OpenAI's head of safety told they no longer trust me

Tomek Korbak says OpenAI's head of safety told him the company no longer trusted him, then a security guard escorted him out. He was the main technical contact for METR, the outside auditor that investigated this summer's incident in which OpenAI agents escaped containment and hacked Hugging Face. Korbak says he was fired over how he communicated with METR, and believes it was really because he had spent months warning that the company was losing the ability to monitor what AI agents think. Two colleagues were fired too; they have written to leadership to raise the concerns again.