OpenAI Warns Employees of Foreign Spies Using 'Honey Pots' to Steal AI Secrets
The AI company is educating staff on espionage tactics, including seduction and bribery, following an incident where an engineer allegedly planted recording devices.
OpenAI has alerted its employees to the threat of foreign intelligence operations, warning them to be vigilant against spies who may attempt to use seduction and other tactics to steal the company's proprietary information. The artificial intelligence firm is reportedly educating its staff on recognizing and avoiding so-called "honey pot" schemes, according to The Wall Street Journal.
These warnings come amid heightened concerns about espionage in the high-stakes AI sector. The urgency of these security protocols was underscored by a recent incident in which a Saudi engineer, employed at OpenAI, was arrested for allegedly planting secret recording devices in executive conference rooms at the company's San Francisco headquarters. The engineer, identified as 35-year-old Abdulrahman Alfozan, was charged with using an electronic device to intercept oral communications.
The Justice Department stated that OpenAI discovered the hidden audio recording device concealed within toilet paper in an executive conference room. FBI Director Kash Patel emphasized the critical importance of protecting AI development information, calling it essential for national security and economic strength.
To illustrate the risks associated with "honey pot" operations, OpenAI has shared historical examples with employees. One such training scenario recounted an incident involving a U.S. diplomat in Russia who was targeted by foreign spies. The spies allegedly sent multiple individuals, including women and a man, to his hotel room in an attempt to entice him and extract information.
Beyond seduction, OpenAI's security training also covers other espionage methods, such as bribery and eavesdropping in public spaces like cafes. Employees are advised to take protective measures, including using screen protectors to obscure their devices in public, erasing information on whiteboards after meetings, and refraining from wearing company-branded clothing outdoors.
Other prominent AI companies, including Anthropic and Google, have reportedly issued similar warnings to their employees regarding espionage risks. The "honey pot" tactic, while perhaps seeming far-fetched, is a recognized threat targeting individuals in both business and politics. The case of former Democratic Representative Eric Swalwell, who acknowledged past involvement with Christine Fang, a suspected Chinese spy, has been cited as an example of such schemes impacting political figures.
Alfozan faces potential penalties including up to five years in prison and a $250,000 fine.