Archive · 9 Oct 2026
OpenAI Fires Three Safety Researchers, Citing Breach of Trust
OpenAI fired three researchers who worked on AI safety.
The company said an investigation found they broke rules about handling sensitive information, but it did not explain exactly what happened.
The researchers said they were worried the firings could make other workers afraid to speak up about safety.
Tomek Korbak said he was told his communication with an outside safety group was related to his dismissal.
Mikita Balesni said he was told he spoke too much with outside safety organisations, but denied leaking company information.
The researchers argued that outside experts can help spot risks in powerful AI systems.
OpenAI said the firings were not because the researchers raised safety concerns.
It said it encourages debate and is arranging work with outside safety assessors.
OpenAI said an internal investigation found Jasmine Wang, Tomek Korbak and Mikita Balesni violated policies on handling sensitive information.
The company called the findings a significant breach of trust but did not disclose details of the alleged violations.
Korbak said he was told his dismissal concerned how he communicated with METR; Balesni said he was told he spoke too much with outside safety organisations.
The researchers warned the firings could discourage employees from raising safety concerns and collaborating with independent experts.
OpenAI denied the dismissals were retaliation for safety criticism and said it was finalizing contracts with third-party safety assessors.
- Who
- OpenAI and former researchers Jasmine Wang, Tomek Korbak and Mikita Balesni.
- What
- OpenAI dismissed the three researchers after an investigation it says found sensitive-information policy violations; the researchers disputed the circumstances and raised concerns about the effect on safety work.
- When
- The dismissals took place the week before the statements and letter published on October 8 and 9, 2026.
- Where
- OpenAI; the articles do not specify where the dismissals took place.
- Why
- OpenAI cited violations of policies on handling sensitive information and a significant breach of trust. The researchers linked their dismissals to communications with outside safety organisations and warned about consequences for safety work.
Researchers' concerns
OpenAI's response
Reason for the dismissals
Researchers' concerns
Korbak said he was told the decision concerned his communications with METR, while Balesni said he was told he spoke too much with outside safety organisations. The researchers disputed the circumstances and raised concerns about the company's priorities.
OpenAI's response
OpenAI said an internal investigation found violations of sensitive-information policies and a significant breach of trust. It did not disclose further details.
Effect on AI safety work
Researchers' concerns
The researchers warned that the dismissals could make staff less willing to raise concerns or work with independent safety experts, and argued that outside collaboration and monitoring are important.
OpenAI's response
OpenAI said the dismissals were not about raising safety concerns, that critical debate is encouraged, and that it is finalizing contracts with third-party safety assessors.
- Researchers dismissed
- Jasmine Wang, Tomek Korbak and Mikita Balesni
- OpenAI's stated reason
- An internal investigation found violations of policies governing sensitive information; OpenAI described the findings as a significant breach of trust.
- Details of alleged violations
- OpenAI did not disclose specific details.
- Korbak's account
- He said he was verbally told the dismissal related to how he communicated with METR during its investigation.
- Balesni's account
- He said OpenAI told him he spoke too much with third-party safety organisations; he denied leaking company intellectual property.
- Researchers' concern
- They said the firings could deter employees from raising safety concerns and working with independent experts.
- OpenAI's response
- The company denied the decisions were about raising safety concerns and said it was finalizing contracts with third-party safety assessors.
Sources
OpenAI fires three researchers over data breach, denies targeting safety critics - Telegraph India
OpenAI defends firing of 3 AI safety researchers, cites ‘significant breach of trust’