A recording device found in an OpenAI office led to an employee’s arrest after the company said its AI models helped trace where the device came from. The employee, 35-year-old Abdulrahman Alfozan, is charged with using or attempting to use a device to intercept oral communications, according to The San Francisco Standard.
What investigators found
The device was discovered plugged into a TV’s USB port in an executive conference room. The San Francisco Standard reported that it was covered in toilet paper and installed on or around September 28th.
OpenAI described it as a “sophisticated recording device.” The company said it used its AI models to examine the device and determine its origin, identifying a specific employee as having planted it a few days earlier.
The company’s account describes AI as part of the effort to understand the device and connect it to a person. It does not provide further details about the models, the analysis they performed, or what evidence was used to reach that identification.
From discovery to arrest
OpenAI said it fired the employee and referred the incident to the FBI. According to The San Francisco Standard, Alfozan was arrested Friday and remained in custody since then, appearing in federal court for the first time on Tuesday.
The FBI investigation is ongoing, and OpenAI said it would continue to support the agency’s work. The charge concerns the alleged use or attempted use of a device to intercept oral communications; the source does not report a finding of guilt.
The sequence described by the company is clear: the device was found, OpenAI used its models to help identify its origin, the employee was terminated, and the matter was passed to federal investigators. The details of the ongoing investigation have not been provided.
What OpenAI says was not accessed
OpenAI said it had no indications that the employee had inappropriately accessed or exfiltrated model weights, source code, or customer data. That statement distinguishes the reported recording-device incident from a known theft or transfer of those materials.
The company also framed the case as an example of AI helping address security threats, including insider risks. That is OpenAI’s description of its role in this incident; the available account does not explain how the models’ findings were independently confirmed or what part they played in the investigation beyond identifying the employee.
Security questions remain open
The discovery in an executive conference room raises questions about recording and privacy in workplace settings. The source establishes where the device was found and how OpenAI says it was concealed, but it does not say whether it successfully recorded a meeting or what, if anything, it captured.
Those unanswered details matter when assessing the incident’s impact. For now, the reported facts are that a device was found, an employee was charged, OpenAI referred the matter to the FBI, and the investigation continues. OpenAI says it has no indication that model weights, source code, or customer data were improperly accessed or taken.