OpenAI has been training employees to watch for spies well before one of its engineers was charged with bugging a conference room, warning them about scenarios like a stranger at a playground asking about their work and agents who use romance to get close, according to The Wall Street Journal. Staff are told not to wear OpenAI-branded clothing outside the office, to put privacy screens on their devices while commuting, and to clean off whiteboards in case cleaning crews have been compromised. Employees at Anthropic and Google have gotten comparable warnings from their own companies.
