Fired over disappearing Signal messages? OpenAI lets go of three outspoken safety researchers

OpenAI has fired three researchers who had spoken publicly about AI safety. The company cited violations of its rules on confidential information; the researchers deny leaking anything.

According to The Wall Street Journal on Oct 22, OpenAI's security team last month called in Tomek Korbak, Mikita Balesni and Jasmine Wang and asked them to unlock their personal phones as part of a leak investigation into a report in The Information about the internals of its new model, Astra. Investigators found that some had disappearing messages turned on in Signal chats with an outside safety researcher. All three were told a week later they were fired.

OpenAI said they were dismissed "for violating clear policies on handling sensitive information," not for raising safety concerns. The researchers say they were never told in writing what they mishandled. In August, they had drafted a proposal urging labs not to develop designs that hide a model's reasoning from humans.

"It sounds like they weren't told not to; they were just fired without warning after the fact. Seems bad," wrote AI researcher Nate Soares, in a post with 560 likes. Balesni wrote that part of him "refuses to believe OpenAI fired three safety researchers for political reasons." Korbak said colleagues are now afraid to talk to outside parties or raise safety issues.