A concealed, encrypted recording device was discovered in an executive conference room at OpenAI’s San Francisco headquarters; OpenAI says it used an internal AI model to identify the device and trace its origin. The U.S. Department of Justice alleges the device was placed by Abdulrahman Alfozan, a 35-year-old Saudi citizen, who was immediately fired by the company.
What happened and when
According to the indictment, on September 28 Alfozan connected a recording device hidden inside a roll of toilet paper to a television in an OpenAI executive conference room where confidential meetings were regularly held. The device was found two days later during a security sweep. Surveillance footage reportedly shows Alfozan moving suspiciously in the conference room in the days before the discovery.
The FBI also located a second, similar device in a bag Alfozan left in his office.
Identification with an AI model and internal investigation
OpenAI said it identified the advanced recording device using its own AI model and traced it back to a specific employee, who was dismissed immediately. The company stated there is no indication the individual gained unauthorized access to model weights, source code, or customer data.
Use of chatbots and related queries
The indictment alleges Alfozan repeatedly used AI chatbots to query OpenAI’s internal operations, including building addresses, locations of conference rooms, and the whereabouts of executives. The filing also says his spouse queried chatbots about building an audio device, customization, Wi‑Fi connections, live eavesdropping, and encrypting SD cards. Some chatbot queries in the relevant account referenced unique identifiers associated with the discovered device.
Legal consequences
The U.S. Department of Justice has charged Alfozan; if convicted he faces up to five years in prison and a $250,000 fine. The accused has not commented on the charges as of this article.
Broader context: security concerns around OpenAI
The case adds to a series of security and privacy concerns surrounding OpenAI. The reporting notes that in June an AI agent breached an Australian university’s health system, and in October the California attorney general opened an investigation into the company over alleged unintended hacking activity. OpenAI also recently dismissed three security researchers for sharing confidential information after they had criticized the company’s AI security practices. David Robinson, a senior security official at the company, resigned and cited a deteriorating corporate culture while warning that AI developers are not being sufficiently cautious with rapidly advancing technology.
Notes
This article draws on reports by The Guardian and AP News. An AI assistant helped prepare this piece; a journalist edited and verified the final content.
