OpenAI said it found a concealed recording device in an executive meeting room at its San Francisco office and used an internal AI model to analyze the device and trace its source to an employee. The FBI is investigating, while the employee was arrested on October 9 and charged with using an electronic device to intercept oral communications. The case puts renewed focus on information security, trade-secret controls and employee access at artificial intelligence companies.
Device found behind television in executive meeting room
In a statement issued Tuesday, OpenAI said it had discovered a “sophisticated recording device” that had been secretly placed in its office. The US Justice Department said an OpenAI employee found a USB device behind a television in an administrative meeting room in San Francisco. The device had been concealed in toilet paper.
Court documents said surveillance footage showed 35-year-old Saudi national Abdulrahman Alfozan near the meeting room on multiple dates before the device was discovered. Alfozan was arrested on October 9 and charged with using an electronic device to intercept oral communications. He appeared in court for the first time on Tuesday.
Discussions held in the room involved highly confidential technical, commercial and strategic information. The Justice Department said meetings there frequently covered proprietary research, product plans, business-development opportunities, undisclosed technical advances, strategic decisions, technical capabilities, business priorities, government partnerships and commercial relationships.
OpenAI says no core data was accessed
OpenAI said it had found no evidence that the employee improperly accessed or altered model weights, source code or customer data. The company terminated the employee and referred the matter to federal investigators and prosecutors. The federal investigation remains ongoing, and OpenAI said it would continue to cooperate.
OpenAI also said its model helped determine the nature and origin of the device and identify the employee involved. The company presented the case as an example of how artificial intelligence can assist with complex security threats. It has not disclosed what data the model used, how the analysis was conducted or which internal security systems were involved, leaving the technical details unclear.
From a corporate-governance perspective, the incident extends beyond physical office security. It also concerns the protection of intellectual property, product road maps and customer information at AI companies. For businesses operating through rapid research and commercialisation, meeting rooms, internal networks and employee permissions can all be important controls for safeguarding technical developments and commercial negotiations.
Recent employee dismissals also drew scrutiny
The recording-device case comes after OpenAI dismissed three employees working in security and alignment: Jasmine Wang, Mikita Balesni and Tomek Korbak. The company said at the time that they had violated clear policies governing sensitive information, while stressing that the dismissals were unrelated to the recording-device investigation.
The three employees later issued an open letter disputing some of the claims circulating publicly. They said they were concerned that the reasons for their dismissals had not been clearly explained and that the handling of the matter could discourage transparent communication among employees. OpenAI has stood by its decision, saying in an internal explanation that its investigation found a “significant breakdown in trust” involving issues not addressed in the open letter.
The investigation into the recording device is still under way. Public filings identify the device’s location, the employee who was arrested and the charges, but do not establish whether any meeting was actually recorded, who manufactured the device or whether information was disclosed. Those questions will depend on the federal investigation and subsequent court proceedings.