OpenAI employee arrested, accused of hiding a recording device in an executive conference room
IA en un minuto newsroom · Editor: Jon Elgezabal
In 30 seconds
An OpenAI engineer was arrested by the FBI on October 9: prosecutors say he hid a covert microphone behind the TV in a room set aside for top management in San Francisco. According to OpenAI, which let him go, nothing so far suggests he took model weights, code or customer data. Kash Patel, head of the FBI, says the alleged aim was to capture talks about superintelligence.
The FBI arrested an OpenAI employee on October 9, accused of hiding a recording device in a meeting room used by management at one of the company's offices in San Francisco. The U.S. Department of Justice and OpenAI itself, which has fired him, made it public on Tuesday, October 13. That same day he made his first appearance before a federal court.
According to the U.S. Attorney's Office for the Northern District of California, the man arrested is Abdulrahman Alfozan, a 35-year-old Saudi citizen who had been working as an engineer at OpenAI since July 2025. He faces a charge of using an electronic device to intercept oral communications. The complaint says OpenAI staff found the device around September 30, wrapped in toilet paper and connected to a USB port at the rear of a TV; the company estimates it was put in place around September 28 and that what it recorded was encrypted.
Prosecutors say Alfozan had no permission to enter that room, that video footage puts him near it on several days before it was found, and that on October 5 he went inside, as reported by CNBC. Around that date, a bag of his turned up in the building holding his wallet, his ID and another, similar device. Before September 28, they add, he had used an internal tool to look up building addresses, rooms, calendars and where management would be. And a chatbot account linked to his wife, which had several identifiers in common with Alfozan's own, was used between September 24 and 30 to ask how to build, install, hide and remotely operate a continuous audio recorder with Wi-Fi. If convicted, he faces a maximum of five years in prison and a fine of up to $250,000.
OpenAI explained on X that it used its own AI models to understand the device and that, so far, it has no sign that this person improperly accessed model weights (the parameters that make a model work), source code or customer data, or took any of them. The company told TechCrunch that the device had advanced encryption, audio listening and a system for moving data out undetected, although the outlet notes it is unclear how it differs from other recorders. Its spokesperson, Nate Evans, confirmed that the employee worked on the product team.
What the recorder was used for is unknown. FBI Director Kash Patel said he allegedly hid it "to secretly capture sensitive communications about the company's super intelligence technology." TechCrunch was unable to reach Alfozan or his lawyer right away.
Why it matters · analysis and opinion
The case shifts where risk sits for AI companies: it is not always an outside cyberattack, but someone with an access badge and a cheap gadget. What exposed it was a mix of ordinary measures, such as cameras, control over who enters each room and logs of internal tool use, rather than some sophisticated defense. For any company with valuable information, AI or not, the lesson is to check the rooms where decisions are made, limit who can enter them and watch for devices plugged into screens or sockets. Still unknown is what the device managed to record, who the man was working for, if anyone, and whether OpenAI will tighten its internal controls after the case.
Source: TechCrunch · Written with the help of AI: how we make the news
