Federal list dismantles OpenAI engineer case by interrogating the smart assistant to uncover meeting rooms

Federal court documents in California revealed details of the pursuit of engineer Abdulrahman Al-Fawzan, 35, after he was accused of attempting to plant covert audio recording devices inside OpenAI's meeting rooms. The official charging indictment, filed by the Federal Bureau of Investigation, states that the engineer, who has been employed by the company since July 2025, faces a charge of using a device or attempting to use it to intercept oral communications, after security staff discovered a sound-pickup wrapped in paper and connected to a television screen inside a room reserved for highly confidential commercial, strategic and governmental discussions.

The case did not involve only the physical surveillance attempt; investigations showed that the defendant also employed the same artificial-intelligence systems to facilitate his movements. Between August and September, Al-Fawzan sent repeated queries to a smart productivity assistant to learn which rooms were allocated for meetings with external partners and company officials, and the daily schedules of those rooms that he was not authorized to enter or view the agenda for. The indictment documents also indicate that in September he used a chat-bot linked to his wife to inquire about how to manufacture, install and remotely access the recording devices, while surveillance cameras recorded his presence in the vicinity of the hall eight times before the device was seized.

Investigators set up an internal trap with a hidden surveillance camera that revealed an attempt to inspect the listening device.After the first device was discovered on September 30, the camera captured Al-Fawzan on October 5 examining the area behind the television, before he left a unmonitored bag in another room that security officials later found to contain a second recording device matching the first, along with his personal ID card and wallet. OpenAI announced that it employed its own artificial-intelligence models to examine the device and determine its origin, confirming that there were no indications of a leak of source code, model weights or customer data, while the company terminated his contract immediately upon his arrest on October 9, leaving him facing a potential penalty of up to five years in prison and a fine of $250,000 if convicted.

In contrast, William Bernick, Al-Fawzan’s defense attorney, affirmed that his client, who has been a lawful permanent resident of the United States since 2020 after completing his studies there, is innocent of these charges until a court proves otherwise. The defense explained that the allegations do not reflect the engineer’s character, urging patience and warning against premature judgments before the investigations are completed and the hearing scheduled for October 19 to consider his continued detention, emphasizing that the indictment does not equate to guilt, and that the investigations have not yet determined the motives for the attempt or the fate of the recordings.

This incident demonstrates that protecting AI secrets extends beyond algorithmic security to safeguarding actual negotiation rooms.At a time when institutions, banks and major regional actors in the Gulf are racing to forge technology and commercial partnerships with advanced model providers, it becomes clear that the details of governmental ties and contractual terms constitute an intelligence target parallel to the models’ own confidentiality. This challenge compels technology and information-security leaders in the region to tighten physical controls over strategic meeting rooms and to audit the logs of queries directed to corporate digital assistants, to prevent daily productivity tools from becoming internal reconnaissance portals that track appointments and executive movements.