OpenAI engineer accused of planting recording device to stay in pre-trial detention

A COURT IN THE United States has granted a request by government prosecutors to prolong the pre-trial detention of Abdulrahman Alfozan, an OpenAI employee, who was arrested by the Federal Bureau of Investigation (FBI) on 9 October 2026. The 35-year-old Saudi national has been charged with attempting to use a recording device to intercept oral communications at the headquarters of OpenAI, where he worked. According to the FBI, Alfozan was attempting to “secretly capture sensitive communications about [OpenAI’s] super intelligence technology”.

Concerns about the engineer’s behavior reportedly started when OpenAI staff discovered a listening device wrapped in toilet paper on 30 September. This prompted an internal investigation, in which OpenAI analyzed the device using its own artificial intelligence (AI) models. The investigation concluded that the device had been installed on or around 28 September. It also determined that the device was highly sophisticated and included advanced encryption, audio monitoring capabilities, as well as a means of transferring intercepted data whilst evading monitoring. Following this discovery, the company referred the case to the FBI.

The affidavit states that Alfozan appears on footage captured on October 5 by a hidden close-circuit surveillance camera at the company’s San Francisco headquarters. The footage shows Alfozan entering a conference room, which he had no authority to access. He then appears to check a device plugged into a television inside the room. The affidavit adds that OpenAI executives frequently use the room for internal leadership meetings, often involving “highly confidential, technical, commercial and strategic information, including proprietary research, product plans, and business development opportunities”.

The complaint also alleges that, prior to installing the device’s, Alfozan used an AI productivity tool to inquire about OpenAI’s buildings, meeting rooms, calendars, and executive locations. It also alleges that a chatbot associated with Alfozan’s spouse—who has not been charged in connection with this case—submitted prompts about building and encrypting an audio recording device.

Alfozan had worked as an engineer at OpenAI for 15 months prior to his termination. According to Al Jazeera, Alfozan’s LinkedIn profile lists earlier roles at Meta and Microsoft. Court filings allege Alfozan had booked one-way flight tickets to Saudi Arabia for himself, his wife, and another relative, scheduled for 12 October. The Saudi national’s connections to the Gulf State’s ruling family are not explicitly known; however, he is allegedly listed as a partner at Raed Ventures, a venture capital firm based in Riyadh. Prince Khalid bin Bader al Saud, a member of the Saudi Royal family, is also a Raed Ventures partner.

Alfozan made his first appearance in a San Franscico federal court on 13 October. He did not enter a plea. If convicted on the charge of using or attempting to use a device to intercept oral communications, he could face a maximum sentence of five years in prison and a $250,000 fine. Notably, Alfozan has not been charged with espionage at this time. OpenAI released a statement saying the company had “no indication of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data”.

This case underscores the increasing threat environment in AI laboratories. As the advent of the AI era raises the strategic value of the technology, so too does the incentive for state and non-state adversaries to target AI firms and tech companies for their trade secrets. Several cases echo Alfozan’s—such as in 2022, when Twitter employee Ahmad Abouammo was convicted of acting as an agent of Saudi Arabia and was sentenced to three and a half years in prison.

Instances of espionage targeting AI and technology companies date back further. In 2001, employees of Sun Microsystems and Transmeta were arrested at San Francisco airport with stolen chip blueprints in their luggage, as they were attempting to board a flight to China. Both pleaded guilty to foreign economic espionage charges.

There has also been an uptick in employees publicly voicing their fears about the lack of security measures and regulations within AI firms. Jacob Coxon, a former OpenAI and Anthropic researcher, resigned from his role in September 2026, and issued a viral statement warning about the safety risks in AI research. He later added that American AI companies had almost certainty been infiltrated by Chinese spies.

With reports that governments such as China and the United States plan to strengthen their AI proficiency, positioning it as a national priority, the global AI race has placed tech firms at the center of state-sponsored espionage. Despite recent attempts to curtail such threats, including training employees to spot espionage attempts, hiring insider risk investigators, and developing rigid security systems, officials at some AI companies have warned that protecting against sophisticated insiders and national state attackers “requires security investments beyond what we’ve currently achieved”.

► Author: Molly Jones | Date: 23 October 2026 | Permalink