U.S. prosecutors have charged OpenAI engineer Abdulrahman Alfozan with allegedly placing a concealed audio-recording device inside an executive conference room at the artificial intelligence company’s San Francisco office, in a case that combines a suspected physical espionage attempt with evidence drawn from office surveillance, workplace records and chatbot prompts.
The Federal Bureau of Investigation arrested Alfozan, 35, on Oct. 9 after OpenAI referred the matter to law enforcement, according to the U.S. Department of Justice. Court filings released on Oct. 13 accuse him of unlawfully intercepting, or attempting to intercept, oral communications through an electronic, mechanical or other device.
Alfozan, a Saudi Arabian citizen who had worked at OpenAI since July 2025, could face up to five years in prison and a $250,000 fine if convicted. He was being held in federal custody ahead of a planned detention hearing on Oct. 19, according to the court record.
Device allegedly hidden behind conference-room television
OpenAI personnel found the first suspected recording device on Sept. 30, prosecutors said. It was plugged into a USB port behind a television in an executive conference room and had been obscured with toilet paper.
Employees estimated that the device had been installed roughly two days earlier, around Sept. 28. The material stored on it was encrypted, the filing said.
The conference room was used for discussions involving confidential technical, commercial and strategic information, according to prosecutors. Topics discussed there included proprietary research, product plans, business-development opportunities, internal technology progress, strategic decisions and cooperation with government entities.
The case does not establish that any meeting audio was successfully captured, transmitted or removed from OpenAI’s systems. The alleged offense covers attempted interception as well as completed interception.
OpenAI told investigators it had found no indication that model weights, source code or customer data had been improperly accessed or taken. Model weights are the numerical parameters that underpin an AI model’s behavior and represent some of the most commercially sensitive assets held by AI developers.
Surveillance footage and searches form prosecution’s case
The charging papers describe several strands of evidence that prosecutors say point to Alfozan.
Investigators reviewed office surveillance footage that allegedly showed Alfozan near the executive conference room repeatedly before the hidden device was discovered. According to the filing, he did not have permission to enter the room or attend meetings held there.
Prosecutors also said that Alfozan returned to the room on Oct. 5 and appeared to inspect the area behind the television where the device had been concealed.
Records reviewed by investigators showed searches for the office address, specific meeting-room details, room capacity and calendars, prosecutors said. The search history also included locations associated with executive gatherings and meetings involving outside partners.
That alleged preparation would place the incident beyond a generic workplace-policy breach. The government’s case focuses on claims that the room was selected because of the confidential subjects discussed inside it and that the device was positioned to collect those conversations.
Chatbot prompts described device construction and encryption
The filing also cites an account linked to Alfozan’s wife that submitted a series of prompts about recording equipment between Sept. 24 and Sept. 30.
According to prosecutors, the prompts covered how to make, customize, install, test and hide recording devices, as well as how to access them remotely. Other questions involved creating a device capable of continuous recording, configuring Wi-Fi connectivity, improving real-time monitoring and encrypting a storage card.
Between Sept. 27 and Sept. 28, the same account asked about identifiers associated with the device, including a custom label and a MAC address, prosecutors said. A MAC address is a hardware identifier used by devices on a local network.
The government has not accused OpenAI’s AI systems of assisting wrongdoing. Rather, the court materials describe chatbot conversations as evidence investigators used to reconstruct alleged planning and technical preparation.
OpenAI separately told investigators that it used an internal AI model during its own review of the device, which it said helped identify Alfozan as the person who installed it. The company did not identify the model or disclose the methods used in that internal analysis.
A second device was found in an office bag
Security personnel found a second, similar recording device on Oct. 5, prosecutors said. It was inside a bag that Alfozan had left in another room at OpenAI’s office, alongside his wallet and identification.
The second device gives prosecutors physical evidence beyond the equipment found behind the conference-room television. Together with surveillance video, search records and chatbot prompts, it forms the core of the government’s allegation that the recording effort was planned rather than accidental.
Prosecutors have not publicly described a motive, identified any alleged recipient of intercepted material or alleged the participation of additional people. The case remains at the charging stage, and Alfozan has not been convicted.
AI companies face growing physical-security pressure
The allegations underline how AI developers’ security risks extend beyond software vulnerabilities and unauthorized access to online systems. A recording device in a room used for product, research and government discussions could potentially expose strategic information without touching a company’s code repositories or internal databases.
For companies developing frontier AI systems, access controls around executive spaces, meeting calendars and internal scheduling tools can become as sensitive as protections around cloud infrastructure. The case also shows how encrypted consumer-grade hardware can complicate an internal inquiry even when the device itself is discovered quickly.
OpenAI’s reported use of an internal model in its investigation adds another dimension. AI tools can help organizations process video, correlate workplace records and identify patterns across large volumes of internal information, though the public filings do not provide enough detail to assess the model’s precise role or performance.
The criminal charge now turns on whether prosecutors can prove that Alfozan knowingly used, or attempted to use, the device to intercept confidential oral communications. The Justice Department’s maximum penalty of five years reflects the seriousness with which federal law treats covert recording in settings where speakers reasonably expect their discussions to remain private.
Concerned about corporate surveillance and crypto safety? Learn essential security practices in crypto safety standards to better protect your assets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
