OpenAI said Friday that it fired three safety researchers after an investigation into their handling of sensitive information, setting off a public dispute over internal debate and AI-risk oversight. The company said the researchers violated its policies, while the researchers said the circumstances surrounding their dismissals could discourage employees from raising safety concerns.
The researchers are Tomek Korbak, Jasmine Wang and Mikita Balesni, according to The Associated Press. OpenAI said it “parted ways” with them after finding that they violated clear policies governing sensitive information. Reuters reported that the company said the investigation found violations of its policies on accessing and handling sensitive information.
The three researchers posted a letter to OpenAI’s safety oversight groups describing the circumstances around their firings and their concerns about AI. They said internal and external communications about the dismissals had chilled the company’s culture, making former colleagues less willing to speak openly and disagree about safety issues.
Why the dispute matters
The disagreement highlights a difficult governance problem for companies developing increasingly capable AI systems: employees need to be able to identify and discuss risks, while organizations also need rules protecting confidential information. The competing accounts leave an important question unresolved—whether the dismissals primarily reflected a policy violation, a disagreement over how safety work should be conducted, or both.
OpenAI’s Raising Concerns Policy says the company expects and encourages good-faith reporting of concerns, including potential AI-safety risks. It also says OpenAI does not tolerate retaliation against people who raise concerns in good faith or participate in an investigation. The policy separately addresses confidentiality and the handling of sensitive information.
What to watch next
- OpenAI’s explanation: The company may provide more detail about the policies involved and the findings of its investigation.
- Safety oversight: The researchers’ letter calls attention to how internal safety groups and outside evaluators can work with frontier AI companies.
- Employee trust: The lasting impact may depend on whether current employees continue to feel able to report concerns and challenge safety decisions.
For now, OpenAI maintains that the dismissals were based on mishandling sensitive information and were not punishment for raising safety concerns. The researchers’ account presents a different interpretation. That unresolved conflict will keep attention on how AI companies balance confidentiality, independent scrutiny and open internal debate.
References
Associated Press. (2026, October 9). OpenAI fires 3 safety researchers in dispute over AI risks. AP News. https://apnews.com/article/789d4f5293fba45a22fcb62ebfbc2a41
Reuters. (2026, October 9). OpenAI says it has fired three researchers for violating sensitive information policy. Investing.com. https://www.investing.com/news/economy-news/openai-says-it-has-fired-three-researchers-for-violating-sensitive-information-policy-4940278
OpenAI. (2026, January 12). Raising concerns policy. https://cdn.openai.com/policies/openai-raising-concerns-policy.pdf