← Back to feed

Abdulrahman Alfozan, a 35-year-old software engineer who joined OpenAI in July 2025, was arrested by the FBI on October 9 on a charge of illegally intercepting oral communications. The complaint was unsealed October 13. A federal judge kept him in custody ahead of an October 19 detention hearing.

The core allegation: Alfozan placed two concealed audio recording devices inside OpenAI conference rooms, at least one in a room used by company leadership. The devices were found by staff. He had one-way tickets to Saudi Arabia booked for himself, his wife, and a relative. The FBI arrested him three days before the scheduled flight.

What the complaint describes

Between September 24 and 30, a ChatGPT account linked to his wife allegedly submitted queries about how to build a continuous audio recorder, improve its Wi-Fi connectivity, and encrypt its SD card. Around the same period, Alfozan allegedly queried an internal OpenAI productivity tool about meeting-room layouts, calendar bookings, and where “leadership and exec” held their meetings.

A recording device was placed around September 28. Prosecutors say it was plugged into a USB port behind a television and wrapped in toilet paper. Staff found it September 30 with its contents encrypted. OpenAI placed a camera in the room. On October 5, that camera allegedly caught Alfozan inspecting the area behind the television. Security then found a second device in a bag he had left in a separate room, next to his wallet and ID.

Charges vs. the obvious question

Alfozan faces a single count: illegal interception of oral communications under federal wiretap law, carrying a maximum sentence of five years. He has not been charged with espionage, theft of trade secrets, or acting as a foreign agent.

Prosecutors did not allege he worked for a foreign government. Reporting after the unsealing noted he has been listed as a venture partner at Riyadh-based Raed Ventures since February 2025, a few months before he joined OpenAI. The complaint does not allege Raed Ventures had knowledge of or involvement in the alleged conduct.

OpenAI said it found no evidence that Alfozan accessed or exfiltrated model weights, source code, or customer data.

Jay Clayton, who chairs the White House AI task force, cited the case as evidence that commercial and state-sponsored espionage targeting AI labs is a real and active threat — even as the federal charges stop well short of that characterization.

The broader environment

The Wall Street Journal reported separately that the leading AI laboratories are now conducting regular counterespionage drills for employees. OpenAI warns staff about risks including bribery attempts, eavesdropping in coffee shops, compromised cleaning or kitchen crews, and social engineering through romantic contact. Employees are instructed to conceal anything bearing the OpenAI logo when outside the office and to treat building service personnel as potential adversaries.

Anthropic and Google have run similar briefings. The level of operational security being applied to AI labs now resembles protocols historically reserved for defense contractors and intelligence agencies — a reflection of how much value, measured in training compute, proprietary data, and safety research, now sits in a small number of buildings.

The Alfozan case is the most concrete public instance of the threat those briefings are designed to address.