OpenAI expands counter-espionage training, tells staff not to wear company-logo clothing outside

OpenAI has rolled out internal counter-espionage training aimed at reducing the risk of employees leaking sensitive business information, according to a Wall Street Journal report cited by BlockBeats. The training uses a spy story as one example, describing how a foreign intelligence agency allegedly tried to seduce a U.S. diplomat in Russia by sending multiple people to his hotel room after an initial refusal. OpenAI used the case to warn staff that romantic or intimate relationships can be exploited to gather confidential information. The company has also trained employees to watch for more ordinary social approaches. In one example, a stranger repeatedly appeared at a San Francisco playground with a child, gradually built familiarity with an OpenAI employee, and later steered casual conversation toward the person’s work. OpenAI has reportedly told staff to avoid wearing clothing with the company logo in public, use privacy screen protectors on laptops while riding the subway, and erase whiteboards after meetings so cleaning staff cannot see sensitive material. The report said these measures were already in place before a bugging case surfaced this week. On Oct. 13, the U.S. Department of Justice said an OpenAI engineer was accused of placing a recording device in an executive meeting room and was arrested by the FBI. Anthropic and Google are also conducting similar training.