Fired OpenAI safety researchers say they were never told what they leaked

Tomek Korbak, Mikita Balesni and Jasmine Wang, three fired OpenAI safety researchers, deny sharing confidential information. They say OpenAI never told them in writing what they allegedly mishandled. They link the firings to their advocacy for AI systems that are easier to scrutinize.

Why it matters: This comes amid several reports about gaps in OpenAI's agent monitoring and its handling of misalignment oversight. If the researchers' account holds, it would suggest internal dissent on oversight is being managed as a confidentiality issue. That bears on how much outsiders can trust a lab's self-reported safety work.