OpenAI trains employees to resist espionage and honeypot tactics

OpenAI has warned employees in training sessions that competitors and state-sponsored actors could try to obtain company information through bribes, eavesdropping in cafes, or seduction, according to a post by Andrew Curran on X. The post says employees at Anthropic and Google have received similar warnings, citing people familiar with the matter. The report does not name the training materials or any specific incident.