Fired OpenAI Researchers Warn Dismissals Could Chill Safety Work
Jasmine Wang, Tomek Korbak and Mikita Balesni dispute OpenAI’s account of their dismissals and say the episode could discourage employees from raising safety concerns.
Topics
News
- Fired OpenAI Researchers Warn Dismissals Could Chill Safety Work
- ARTEX Developer Closes AI Hacking Tool After Bank Attacks
- AI Redraws Global Race for Quant Talent as Coding Tools Lower Entry Barriers
- AI Dispatch | India Draws Billions Into AI
- AI Bots Now Make Up Half of Traffic to US Fed’s FRED Database
- Musk Blames ‘Oligarchs’ For Delay in Starlink’s India launch
Three former OpenAI safety researchers have challenged the company’s account of their dismissals, warning that the firings could make employees less willing to raise concerns or work with outside safety organizations.
Jasmine Wang, Tomek Korbak and Mikita Balesni published a joint open letter on Thursday addressed to OpenAI’s Safety and Security Committee, Safety Advisory Group and Mission Advisory Council.
“We have become concerned that internal and external communications around our firing have made our former colleagues afraid to speak and operate in ways that, until last week, were an integral part of working at OpenAI,” they wrote in the letter.
The researchers said OpenAI had previously encouraged employees to raise safety concerns and collaborate with outside experts, particularly where independent evaluation was needed.
“AI is not a normal technology, and OpenAI is not a normal company,” they wrote. “Those of us who work on safety see risks before anyone else, and we rely on close collaboration with outside experts to work out how to address them.”
OpenAI said the dismissals followed an investigation into the handling of sensitive research information.
A company spokesperson told TechCrunch that the inquiry found a “pattern of misconduct” that went beyond information shared with an external AI evaluation organization.
OpenAI has not publicly detailed all of the alleged policy violations.
The three researchers deny that they acted outside the scope of their work. Their letter also rejects any involvement in a leak to The Information concerning changes that could make the reasoning of newer OpenAI models harder to monitor.
Balesni said his work on model monitorability required extensive engagement with outside parties and that he coordinated with his reporting line, research leadership and board members. The letter says he removed sensitive details before sharing material externally and acted within company norms as he understood them.
Wang separately described her own dismissal in an X thread, saying OpenAI told her she had been fired for accessing an executive’s email.
“OpenAI delegated that access to me for recruiting,” Wang wrote. She said she had asked IT to remove the access, later opened a sensitive email unintentionally and informed the executive within minutes.
“The reasons that we were provided for our terminations are simply not adding up,” she wrote.
Balesni also shared the full letter on X, while Korbak said publicly that he had been told his dismissal related to the way he communicated with external evaluator METR, work he said formed part of his role.
OpenAI has rejected the suggestion that the firings were retaliation for safety advocacy. An internal memo shared with TechCrunch said the decisions “were not about raising safety concerns or speaking out” and that the company continues to encourage employees to do so.
The researchers are asking OpenAI to maintain independent third-party safety evaluation, preserve the monitorability of frontier models and set clearer rules governing contact between employees and outside safety researchers.
