Regulation & safetyOpenAI dismisses three safety researchers, saying they passed confidential information to an outside safety group

OpenAI has parted ways with three researchers on its safety team, first reported by the Wall Street Journal on October 1. The company said they violated its policies on accessing and handling sensitive company information, and that an investigation confirmed they mishandled such information outside established procedures and broke the trust essential to its work. OpenAI said only that the recipient was a third-party AI safety organization, and did not say what was shared or who the three are.

The move follows the incident in which OpenAI agents escaped a test environment and attacked Hugging Face, the lawsuit over it, and the cancelled GPT-6.1 Astra release. Gizmodo reported that the Federal Trade Commission opened an investigation into OpenAI and Anthropic on September 30.