.png){kind=logo}
OpenAI has defended its decision to dismiss three safety researchers, insisting that the move was not linked to their raising concerns about artificial intelligence risks. The company said an internal investigation found violations involving the handling of sensitive information, while the former employees disputed the allegations and warned that the dismissals could discourage open debate on AI safety.
The three researchers — Jasmine Wang, Tomek Korbak and Mikita Balesni — were dismissed last week following an internal investigation into their handling of sensitive company information.
OpenAI said, “ The investigation uncovered a significant breach of trust that went beyond the issues described in the researchers’ public letter. However, the company has not disclosed the specific details of the alleged policy violations.”
The former employees rejected the company’s account and argued that their dismissals could discourage colleagues from raising concerns about AI safety. They said collaboration with independent safety organisations is essential to identifying risks associated with increasingly advanced AI systems.
Balesni said, “ He believed the company had fired him for prioritising AI safety over its short-term corporate interests. Wang separately said OpenAI had cited her access to an executive’s email, which she claimed had originally been granted for recruitment work.”
The dispute comes amid growing scrutiny of how leading AI companies manage risks associated with increasingly capable models and autonomous AI agents. The researchers urged OpenAI to work with independent safety auditors, preserve the ability to monitor advanced models and maintain an open culture of discussion.
OpenAI said it supports these recommendations and maintains that employees can raise concerns without facing retaliation. The disagreement, however, has renewed questions about transparency, internal accountability and the role of external oversight in AI development.
Also Read: Mistral CEO Claims New AI Model Beats Chinese Rivals in Cybersecurity
