OpenAI Fired Us for Safety Concerns, Researchers Claim — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Mikita Balesni, Tomek Korbak, and Jasmine Wang — all former safety or alignment employees at OpenAI — posted an open letter on X stating they were fired for "prioritising safety over the near-term interests of OpenAI as a corporation."
- OpenAI's spokesperson said on 2 October that its investigation "confirmed that these individuals mishandled sensitive information outside established company procedures," and research leaders reiterated in a Friday note that "these decisions were not about raising safety concerns or speaking out."
- OpenAI's Friday note cited "a significant breach of trust beyond what's outlined in the letter," while Tomek Korbak separately said he had been "raising safety concerns that we're losing the ability to monitor what AI agents think" and believed that was why he was fired.
- Jasmine Wang warned that the former employees "were not the first to be pushed out of OpenAI under suspicious circumstances" and she was "concerned we will not be the last" unless staff take a stand.
- OpenAI said it is finalising contracts with third-party safety assessors and will announce details in the coming weeks — a step the fired researchers said has left remaining staff "afraid to speak" as before.
- Alignment work — the practice of embedding human ethical principles into AI systems — is what all three researchers said they were doing, and their joint letter framed the dispute as part of a "growing debate over AI safety and the potential risks advanced AI systems could pose to society."
Why it matters: One fired researcher warned this firing pattern means remaining employees "will not be the last" unless they resist, which escalates a single HR case into a public referendum on OpenAI's internal safety culture. For outside regulators, journalists, and customers, the dispute forces a credibility call: did three alignment workers mishandle information, or did OpenAI reclassify safety whistleblowing as a "breach of trust"?
Ask SkimNews
