OpenAI Upholds Firings of Three AI Safety Researchers

What changed

OpenAI is standing by its decision to fire safety researchers Jasmine Wang, Tomek Korbak and Mikita Balesni.

The company says an internal investigation found a “significant breach of trust” and violations of policies governing sensitive information. The researchers dispute that explanation, saying they were dismissed after raising concerns about AI safety. OpenAI has not disclosed what conduct or information the investigation involved.

Why it matters

This is now a dispute over both conduct and credibility. OpenAI says the firings were not retaliation for speaking about safety. The researchers say they acted within the company’s mission and workplace norms.

That unresolved gap matters inside an AI lab building increasingly capable systems. Safety researchers need clear rules for handling sensitive information, but they also need to know where raising concerns ends and a firing offence begins. When the company gives only the label “breach of trust,” colleagues are left to judge a serious employment decision without the underlying facts.

In the short term, the dispute may make internal criticism and safety disclosures look riskier to OpenAI’s researchers. Over the next six to twelve months, other AI companies hiring frontier-model safety staff may face sharper questions about confidentiality rules, investigation procedures and protections for dissent. The answer will depend partly on whether OpenAI provides more detail about what happened and how its policies were applied.