Federal authorities arrested an OpenAI engineer after the company uncovered a concealed recording apparatus in a high-security conference room. The individual, identified as Abdulrahman Alfozan, a 35-year-old Saudi national, faces a federal charge of using or attempting to use a device to intercept oral communications, according to a Department of Justice release. The arrest occurred on Friday and the suspect has remained in custody pending a court appearance.
OpenAI’s internal security team discovered the device plugged into a television’s USB port within an executive meeting space where highly confidential technical, commercial and strategic discussions are routinely held. The apparatus was wrapped in toilet paper and its storage was encrypted, suggesting deliberate concealment. Investigators determined the installation took place around September 28, and the device’s presence was first noted during a routine inspection of the room’s equipment.
CCTV footage reviewed by the Department of Justice shows Alfozan near the executive conference room on multiple occasions before the device’s detection. The video captures him appearing to peer behind the television where the recorder was later found. In a separate office, security staff recovered a bag left by the employee that contained his wallet, identification and a second recording device, further corroborating the alleged insider operation.
OpenAI immediately terminated Alfozan’s employment and referred the matter to the FBI, which has taken custody of the suspect. In a public statement, the company said it used its own AI models to analyze the device and trace its origin, expressing gratitude to federal agents for a swift arrest. OpenAI emphasized that its security program is designed to counter a range of threats, from insider actors to nation-state adversaries.
The DOJ charge cites a violation of federal wire-tap statutes, specifically the unlawful use or attempted use of a device to intercept oral communications. Prosecutors allege that Alfozan deliberately placed the recorder to capture privileged conversations, a breach that could expose proprietary AI research and business strategies. The case proceeds under federal jurisdiction, with potential penalties reflecting the seriousness of insider espionage.
OpenAI reported no evidence that the suspect accessed or exfiltrated model weights, source code or customer data, suggesting the intrusion was limited to conversational capture. The incident highlights growing concerns about insider threats within AI firms and the role of advanced detection tools. By leveraging its own models to dissect the device, OpenAI demonstrates a proactive approach to safeguarding intellectual property and operational secrecy in a competitive industry.