OpenAI Engineer Charged Over Alleged Office Recording Device

Alleged surveillance at an AI company office

Federal prosecutors have charged an OpenAI engineer in connection with an alleged attempt to secretly record conversations in an executive conference room at one of the company’s San Francisco offices, according to a criminal complaint described by Wired.

The complaint alleges that Abdulrahman Alfozan, identified by prosecutors as a 35-year-old Saudi Arabian national and United States permanent resident, hid an audio-recording device behind a monitor in a conference room used by senior executives. Prosecutors allege the device was concealed in toilet paper and inserted into a USB port on the back of a television.

Alfozan made an initial appearance in federal court in San Francisco on one count of using an electronic, mechanical, or other device to intercept oral communications. A conviction could carry a sentence of up to five years in prison. The allegation is not a finding of guilt, and the case will proceed through the courts.

What prosecutors allege

According to the complaint, the alleged operation was not limited to placing a device in a room. Prosecutors say Alfozan spent several weeks using an unspecified productivity agent and an AI chatbot while planning the alleged surveillance. He also allegedly used an internal OpenAI tool to ask about meetings scheduled in the conference room that was targeted.

The complaint further alleges that a chatbot account associated with Alfozan’s spouse was used to ask technical questions about setting up a covert audio-recording device. Wired reported that OpenAI’s security team also found another recording device in Alfozan’s bag at the office.

The company discovered the device on September 30 during a countersurveillance sweep, according to the report. Alfozan joined OpenAI in July 2025 and was fired the Friday before his arrest, Wired reported.

A security incident without a reported data breach

The allegations arrive at a particularly sensitive moment for companies developing advanced AI systems. Conference rooms used by executives can contain strategy discussions, product planning, security decisions, and commercially sensitive information. At an AI developer, that can include conversations about highly valuable research and systems.

FBI Director Kash Patel said in a statement that Alfozan sought to secretly capture sensitive communications about the company’s superintelligence technology. Prosecutors and OpenAI had not publicly offered information establishing a motive, according to Wired.

OpenAI said it had no indication that the individual inappropriately accessed or exfiltrated model weights, source code, or customer data. In a statement posted on X, the company said it takes the protection of its technology, data, and people seriously.

That distinction matters. The incident, as described, concerns alleged interception of conversations rather than a confirmed compromise of the company’s core technical assets. It does not mean the alleged conduct was minor: recordings of internal deliberations can themselves expose sensitive information. But OpenAI’s statement narrows what has been publicly reported so far about the scope of any potential harm.

AI tools in the alleged planning

The complaint’s references to an AI chatbot and a productivity agent are likely to draw attention because they place familiar workplace tools inside an alleged criminal investigation. The available account does not say that an AI system carried out the alleged surveillance or that OpenAI’s models were compromised. Rather, prosecutors allege that conversational and internal tools were queried during planning.

That is an important difference. General-purpose software can be used for benign research, work planning, and technical questions, but its availability does not eliminate the need for organizations to protect sensitive physical spaces. The alleged incident illustrates how workplace security now spans digital systems, meeting schedules, access controls, hardware inspections, and employee conduct.

It also underscores a basic limitation of cybersecurity alone. A company can secure networks, repositories, and devices while still facing risk from a small physical recorder placed in a room. Regular physical sweeps, reporting procedures, and clear access controls remain part of protecting information that is spoken rather than stored in a database.

What remains unresolved

Several key facts remain unclear in the public reporting. Prosecutors and OpenAI had not disclosed a motive. There was no public indication that any recordings were successfully captured, retrieved, or shared. The reported allegations also do not establish what information, if any, may have been exposed.

Alfozan could not immediately be reached for comment, Wired reported. The case will determine whether the government can prove its allegations. Until then, the most firmly established public facts are the criminal charge, the reported discovery of devices, and OpenAI’s statement that it has not found signs of inappropriate access to or exfiltration of model weights, source code, or customer data.

A broader lesson for high-value research organizations

For AI companies and other organizations handling sensitive research, the episode is a reminder that information security includes the rooms where people meet. Protecting advanced technology requires more than guarding code and servers. It also requires procedures that account for the human, physical, and operational pathways through which confidential discussions can be exposed.

The allegations will now be tested in court. In the meantime, OpenAI’s reported response indicates that its internal investigation focused both on securing the workplace and determining whether any of its most sensitive technical assets had been accessed or removed.

This article is based on reporting by Wired. Read the original article.