OpenAI fires three safety researchers over sensitive information
OpenAI says it parted ways with three researchers for mishandling sensitive information, which the Wall Street Journal reports they shared with an outside AI-safety organisation.
OpenAI said it has "parted ways with three individuals" for violating its policies on handling sensitive company information. The Wall Street Journal reported, citing people familiar with the matter, that the three worked on OpenAI's safety team. They allegedly passed confidential information to an outside AI-safety organisation.
An OpenAI spokesperson said an internal investigation confirmed the three "mishandled sensitive information outside established company procedures." The spokesperson said that broke company policy. In the reports we read, OpenAI has not named the researchers. It has not named the organisation said to have received the information either.
On X, reaction ran ahead of the facts. Andrew Curran wrote that three alignment researchers leaving that morning had already been discussed on the platform. His post had 541 likes when we read it. Nathan Calvin wrote that two researchers he named led a chain-of-thought monitorability paper, and that their exit could be terrible if the Journal story refers to them.
Calvin's identification is his own inference. The Journal's account, as relayed by others, does not name anyone, and OpenAI has not confirmed who left. The firings were reported the same day California's attorney general served OpenAI with a subpoena over cybersecurity incidents. Nothing we have seen links the two.