Someone hid a listening device inside a roll of toilet paper, then plugged it into a USB port on the television in OpenAI's executive conference room in San Francisco. The bug sat there long enough to catch conversations the company says touched on its most sensitive work: research toward superintelligence.
The FBI has arrested Abdulrahman Alfozan, a 35-year-old Saudi national who had worked at OpenAI as an engineer since July 2025. He's charged with using a device to intercept oral communications, a crime that carries up to five years in prison and a $250,000 fine. He made his first court appearance Tuesday, October 13; the next hearing is set for October 19.
Investigators say the device went in around September 28 and was found just two days later. OpenAI's own AI models reportedly helped crack the case — the company ran an internal analysis on the hardware that led straight back to Alfozan. Then, on October 5, staff found an unattended bag in another building containing his wallet, his ID, and a second, nearly identical recording device.
OpenAI says it found no sign the engineer accessed or copied model weights, source code, or customer data — so the company's core technology appears untouched. Still, FBI Director Kash Patel described the device as an attempt to secretly capture "sensitive communications about the company's superintelligence technology," and the Justice Department is treating the case as a national-security matter. OpenAI fired Alfozan the same Friday he was arrested.
One detail adds an edge to the story: Alfozan is also a venture partner at Raed Ventures, a Saudi firm whose partners include a member of the kingdom's ruling family. Nothing in the charges suggests the fund knew about the bug or had any role in it. But an AI lab engineer with ties to foreign capital bugging the room where the most expensive secrets get discussed is exactly the kind of scenario pushing frontier labs to rethink how they vet staff before handing them access to that room.
