Recording Devices Found in OpenAI Offices

The FBI arrested Abdulrahman Alfozan on Oct. 9 after investigators linked him to recording devices found in OpenAI offices. OpenAI said it fired him that day. The investigation centers on a device found in a conference room used for highly sensitive discussions and another in a bag he had left in a separate room. The case has raised concerns about security inside frontier artificial intelligence laboratories.

Devices in Sensitive Rooms

The Justice Department said OpenAI's security team found one device late last month during a sweep of a conference room. It appeared to sell for about $14. Investigators said someone had plugged it into a television's USB port and wrapped it in toilet paper.

Investigators traced the device to Alfozan, an engineer who joined OpenAI last year. Surveillance video showed him near the conference room around the time investigators suspected someone had installed it. The Justice Department said he didn't have authority to use the room.

Investigators also found another recording device in a bag Alfozan had left in a separate room. The source account doesn't describe what that second device recorded or whether investigators recovered recordings from it. OpenAI said it used its models to help study data copied from the device.

The Investigation and Allegations

Investigators found queries about building a continuously recording device and encrypting its data on a chatbot account associated with Alfozan's spouse. The account says investigators found the queries during the investigation, but it doesn't state that the spouse made them or knew about the devices.

The Justice Department described Alfozan as a 35-year-old Saudi national. He worked on OpenAI's agent software and previously worked at Meta, according to a LinkedIn profile that appeared to be his. He was also a part-time partner at a Saudi-based venture capital firm.

The arrest is an allegation under investigation, not a conviction. If convicted, Alfozan could face up to five years in prison, a $250,000 fine and three years of supervised release. The source account doesn't identify a motive or say investigators have established that anyone passed information to another person or organization.

Security Concerns at AI Laboratories

According to the report, the discovery has prompted concern about the possibility of spies inside frontier AI laboratories. Jacob Coxon, a former Anthropic researcher who warned about artificial intelligence risks, had raised alarms about that possibility at Anthropic and OpenAI.

The report said the incident underscored the potential reality of the threat. It describes the company security team's sweep, investigators' use of surveillance video and the devices found in two rooms, but doesn't detail any changes OpenAI has made to its security procedures.

The investigation concerns both the specific devices and access to a room reserved for sensitive discussions. OpenAI said it fired Alfozan on Oct. 9, the same day the FBI arrested him. The available account doesn't say what information, if any, the devices captured before investigators found them.