OpenAI Terminates Three Security Researchers Over Probe Into Signal Chat and Model Leak

According to The Wall Street Journal, OpenAI terminated three security researchers. Internal security personnel had previously discovered that they were engaging in disappearing-message chats on Signal with external AI security researchers, raising suspicions that sensitive information, such as details about OpenAI’s internal model designs, may have been shared. The investigation reportedly began following a September article about OpenAI’s latest model, Astra. The article mentioned that Astra employs a novel design that enhances model capabilities while reducing the observability of its reasoning processes by humans. OpenAI viewed the report as a potential leak of internal model design details and initiated an investigation to trace the source.