OpenAI fires three security researchers over investigation into Signal chats and model information leaks

According to The Wall Street Journal, OpenAI fired three security researchers. Internal security staff previously discovered that they had been exchanging disappearing messages via Signal with external AI safety researchers, and suspected that the chats contained sensitive information, including details about OpenAI’s internal model design. The investigation reportedly began following a September article on OpenAI’s latest model, Astra. The report noted that Astra employs a new design that enhances model capabilities while reducing human-observable reasoning processes. OpenAI viewed the article as a potential leak of internal model design and began tracing the source.