Key points
- Abdulrahman Alfozan is charged with using a device to intercept oral communications.
- Prosecutors allege he used AI chatbots to research surveillance techniques and meeting schedules.
- OpenAI security teams found the device during a countersurveillance sweep on September 30.
The short version
- An engineer at OpenAI allegedly hid an audio recording device behind a monitor in an executive meeting room.
- Federal prosecutors state he used internal tools and external chatbots to plan the surveillance operation.
- The company fired him and confirmed no model weights or customer data were exfiltrated.
The details
According to a criminal complaint unsealed on Tuesday, Abdulrahman Alfozan concealed an audio recording device in toilet paper and attached it to a monitor in an OpenAI conference room in San Francisco. Federal prosecutors describe Alfozan as a 35-year-old Saudi Arabian national and US permanent resident. He joined the company in July 2025. He was fired last Friday, coinciding with his arrest.
He made an initial court appearance in San Francisco on Tuesday. He faces one charge of using a device to intercept oral communications. A conviction could result in up to five years in prison. OpenAI’s security team discovered the device on September 30 during a countersurveillance sweep. The device was allegedly inserted into a USB port on the back of a television. OpenAI also found an additional recording device in Alfozan’s bag at the office.
The context
Prosecutors allege Alfozan spent several weeks querying an unspecified productivity agent and an AI chatbot to plan the surveillance. He used an internal OpenAI tool to check meeting schedules for the targeted room. He also used a chatbot account linked to his spouse to ask technical questions about setting up covert audio recording devices.
FBI Director Kash Patel stated in a release that Alfozan aimed to capture sensitive communications about the company’s super intelligence technology. However, neither prosecutors nor OpenAI have provided information regarding a specific motive. OpenAI stated on X that there are no indications he accessed model weights, source code, or customer data. The company emphasized it takes protecting technology and data seriously.
What to watch
Alfozan previously worked at a cargo logistics startup and at Meta, according to a purported LinkedIn profile. He also listed himself as a part-time venture partner at RAED Ventures, a Saudi Arabian firm. In a recent podcast interview, he said he was part of a group of Saudi engineers offering consulting services to Saudi organizations. OpenAI has not confirmed if other devices were planted or if the chatbot queries were monitored in real time. The criminal complaint does not detail how the countersurveillance sweep was triggered. No other employees have been named in connection with the incident.
What it means for organisations
Teams usually need to review access logs for internal tools that reveal schedule or location data. Check whether chatbot usage policies prohibit technical queries related to physical security or surveillance. Security leaders should assess if physical countersurveillance sweeps are scheduled regularly or only after an incident. Communicate clearly with legal counsel about potential insider threats. Ensure that personnel with access to executive spaces undergo appropriate background checks. Monitor for unusual patterns in tool usage that might indicate planning for unauthorized activities.
What to do and how to stay safe: OpenAI
- Review internal tool access logs for queries that reveal sensitive meeting schedules or locations.
- Audit chatbot usage policies to ensure they prohibit technical questions about surveillance or physical security breaches.
- Conduct regular physical countersurveillance sweeps in executive meeting rooms and high-security areas.
- Verify that all personnel with access to sensitive spaces have undergone recent background checks.
Step-by-step guide: Secure Home Routers for Small Business Networks
General security guidance from the TechSec Headlines newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Did the engineer steal any data from OpenAI?
OpenAI stated there are no indications the individual accessed or exfiltrated model weights, source code, or customer data.
What charge is the engineer facing?
He faces one charge of using an electronic or mechanical device to intercept oral communications.
How was the recording device discovered?
OpenAI security teams found the device during a countersurveillance sweep on September 30.
