In brief
Three former OpenAI researchers were let go amid an internal investigation that the company says uncovered mis‑handling of sensitive data. The ex‑staff, all safety and alignment specialists, counter‑claimed they were dismissed for speaking out about AI safety risks. A joint open letter decries a culture that now silences concerns. OpenAI maintains that the firings were justified on internal policy grounds and vows to bring in third‑party safety assessors. The dispute highlights the growing tension between corporate agendas and ethical oversight in AI development.
Three former members of OpenAI’s safety team say they were fired because they voiced worries about the risks of advanced AI, while the company insists the decisions were based on violations of its data‑handling policies.
On 2 October, a spokesperson told the BBC that an internal probe found the researchers had improperly shared confidential information outside company protocols. The ex‑employees responded on X with an open letter that they believed the dismissals were retaliation for raising safety concerns about the company’s AI agents.
OpenAI’s research leaders issued a statement saying, “Our investigation uncovered a significant breach of trust beyond the letter published in the open. We stand by the decision to not continue their employment.” The note rejected the claim that the firings were tied to safety advocacy.
The former staffers – Mikita Balesni, Tomek Korbak and Jasmine Wang – highlighted that their removal has left remaining employees uneasy about speaking publicly on safety topics. Korbak posted that he had long worried about losing tools to monitor AI agent behaviour, a concern he believes was the real cause of his dismissal.
OpenAI said it is finalising contracts with third‑party safety assessors and will announce the arrangements in the coming weeks, a move that may be aimed at restoring confidence among staff and investors after the controversy.
