A Saudi national has been arrested on suspicion of bugging OpenAI’s headquarters following an FBI investigation.

Abdulrahman Alfozan, 35, was an employee of the tech giant and was arrested on Friday over claims he had planted a device at the group’s San Francisco base.

According to prosecutors, he bugged the company by plugging a device into a USB port on the back of a television in an executive conference room, which had been covered by toilet paper.

He has been charged with using or attempting to use a device to intercept oral communications, US prosecutors confirmed on Tuesday.

The Saudi national worked at OpenAI, which is behind ChatGPT, since July 2025 as an engineer.

Employees had discovered the device on Sept 30 and determined it had been planted two days earlier.

Surveillance footage allegedly showed Mr Alfozan near the room on multiple dates before the discovery, despite him not being cleared to enter the room or attend meetings there, prosecutors said.

Later, OpenAI workers found a similar device on Oct 5 in a bag found in the company’s offices containing Mr Alfozan’s ID and wallet.

The 35-year-old made an initial appearance in federal court on Tuesday and remains in police custody.

OpenAI uses AI to ‘understand’ bug device

In a statement shared on X, OpenAI said it had discovered a “sophisticated recording device” hidden inside one of its offices and had fired Mr Alfozan.

It said it used its own AI models “to understand the device and where it had come from”, which led to identifying that a specific employee had planted it a few days earlier.

The company then referred the matter to the FBI.

It said there were “no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data”.

OpenAI said it treated protecting its technology, data, and people “seriously, and as a matter of national importance”.

According to the complaint, meetings in the room reportedly involved confidential technical and strategic information, including product plans and government engagements.

Mr Alfozan allegedly used a productivity tool to ask about OpenAI’s operational details including building addresses and meeting rooms, as well as the locations of leadership and outside partner meetings.

Craig Missakian, the US attorney for northern California, said that the issue “is a matter of national security”.

Mr Alfozan could face up to five years in prison and a $250,000 fine, followed by three years of supervised release, if found guilty. He is yet to enter a plea and prosecutors said under US law he is innocent until proven guilty.

A detention hearing before Alex Tse, a US magistrate judge, is scheduled for Oct 19.