The FBI has arrested Abdulrahman Alfozan, a 35-year-old OpenAI engineer, with charges announced on October 13, 2026. He is accused of planting a thumb-sized recording device concealed in toilet paper and plugged into a conference-room TV's USB port, where it was discovered on September 30.

According to reporting from USA Today, OpenAI used its own models to help identify the engineer behind the device. The company has since terminated him and says no model or customer data was exfiltrated — though with charges only just announced, the case details are still unfolding.

The story reads like a cautionary tale about insider threats at the world's most scrutinized AI labs. OpenAI sits at the center of geopolitical and commercial interest in AI capabilities, and even the allegation of a planted recording device raises uncomfortable questions about how secure internal spaces really are at frontier labs.

It also highlights a wider industry pattern: as AI labs become more secretive about models, training data, and internal research, they become more attractive targets — not just for outside hackers, but for people on the inside. Physical security, not just cybersecurity, is part of the equation.

Worth remembering: Alfozan has been charged, not convicted, and the specifics of what the device captured — or was intended to capture — have not been made public. We'll update this story as official details emerge.

Sources