OpenAI, Anthropic and Google Warn Staff of Espionage Risks as AI Secrets Attract Spies

Espionage Warnings Across Silicon Valley

AI companies are warning employees about growing espionage risks as valuable technology and trade secrets attract foreign agents, competitors and other bad actors. OpenAI has held training sessions warning that outsiders, including competitors and state-sponsored actors, could try to obtain company information by offering money, listening in cafes or even attempting to seduce staff. Employees at Anthropic and Google have received similar warnings.

One scenario presented to OpenAI employees involved a stranger at a San Francisco playground who repeatedly befriended an employee there with their child and asked about their work. The stranger was a spy. Trainings also cover so called honey pots, in which agents use the offer of sex or romance to get close to a target. Employees are told to avoid wearing OpenAI branded clothing outside the office, use screen protectors on public transit and wipe down whiteboards.

OpenAI Employee Charged in Recording Case

On Tuesday, a Saudi employee of OpenAI, Abdulrahman Alfozan, was charged with trying to record audio using a device planted behind a TV set in a company conference room. FBI Director Kash Patel said protecting the technological secrets behind AI is essential to safeguarding the nation's economic strength and national security, and that the FBI views that mission as one of its highest counterintelligence priorities.

Alfozan's attorney said the allegations were wholly out of character and noted his client was charged only with violating a federal recording statute, not with stealing trade secrets or acting as a foreign agent. Corporate security experts say the threat likely extends well beyond one hidden recording device. Randy Watt, president of consulting firm SRW, said catching one guy means there are half a dozen out there, and that bad actors may seek personal secrets to exploit or blackmail individuals.

Insider Risks Grow With Rapid Hiring

The rapid growth of the industry has created abundant opportunities to get inside leading companies, which have quickly added thousands of employees. OpenAI, Anthropic, Meta and Google all employ large numbers of foreign nationals, with China supplying much of the world's top AI talent. Trump administration officials have identified the employment of Chinese nationals by AI companies as a national security vulnerability, alleging that some share secrets with Beijing.

Employees also frequently job hop between companies, creating an additional vector for information leakage. OpenAI and Apple are engaged in litigation over claims of poached employees taking confidential files to new jobs, and Elon Musk's xAI made a similar claim against an engineer who went to OpenAI. Aviv Nahum, chief executive of Above Security, said AI companies are scaling workforces at unprecedented speed, and the question is whether insider risk management is scaling at the same pace.

High Stakes Make AI Firms Prime Targets

Phil Venables, a partner at Ballistic Ventures and former Google Cloud chief information security officer, said corporate espionage has existed as long as there has been intellectual property to steal. In 2022, a former Twitter employee was sentenced to 42 months in prison for spying for Saudi Arabia. In January, a former Google software engineer was convicted of stealing thousands of pages of confidential AI trade secrets information and giving the material to China. Venables said the concentration of value in a handful of AI companies makes them bigger targets.

Sources: Livemint · Wall Street Journal · Techmeme