OpenAI has announced that it has severed ties with three researchers for violating policies regarding access to and handling of sensitive corporate information. The company's investigation confirmed that these individuals inappropriately handled confidential information without following established internal procedures, undermining the trust essential to their operations.
SecurityOpenAITomas CorbachoJasmin WanMikita Balesh
OpenAI Dismisses Three Researchers Over Allegations of Sharing Confidential Information with External AI Safety Organizations
This article is a translation. Read the Japanese original
According to a report by the Wall Street Journal citing sources familiar with the matter, the three dismissed individuals are Tomek Kolbak, Jasmine Wang, and Mikita Baleshni, all members of OpenAI's safety team. Mr. Kolbak was responsible for technical communications with external evaluation organizations, including METR and Redwood Research.
OpenAI is currently facing pressure from external organizations to undergo safety testing by third-party agencies following an incident where its AI mistakenly hacked Hugging Face, and it is currently undergoing investigations by staff from organizations such as METR.
Sources
- OpenAIが機密情報の取扱ポリシー違反で研究者3人を解雇、AI安全組織に情報を共有したため （GIGAZINE、2026-10-02）