Three former researchers at OpenAI have claimed their dismissal was linked to raising safety concerns about the company’s AI systems, sparking a controversy over how the organisation balances innovation with risk management.
“I believe we were fired for prioritising safety over the near‑term interests of OpenAI as a corporation,” wrote Miak Balesni on the social‑media platform X, accompanied by an open letter signed by his colleagues, Tomas Korbak and Jasmine Wang.
OpenAI’s spokesperson, speaking on 2 October, said an investigation “confirmed that these individuals mishandled sensitive information outside established company procedures,” and reaffirmed the “significant breach of trust” that prompted the firings.
The letter’s authors are all former members of the firm’s safety or alignment teams—groups responsible for embedding human values into AI behaviour and monitoring agents for misbehaviour.
“For months, I’d been raising safety concerns that we are losing the ability to monitor what AI agents think, one of our best tools for catching when they misbehave,” Korbak wrote, suggesting that such concerns may have been the real reason for his termination.
OpenAI said it is finalising contracts with third‑party safety assessors and plans to reveal more details in the coming weeks, but its statement maintains that the dismissals were not about whistleblowing.
The dispute comes amid an industry‑wide debate on the risks posed by advanced AI, with critics arguing that a culture of fear could stifle essential safety oversight.
The researchers’ claims may prompt broader scrutiny of how leading AI labs manage internal safety discussions and whether their employees feel safe to raise concerns before policy changes are implemented.
