A Saudi engineer allegedly hid an encrypted listening device behind a TV in OpenAI's executive conference room. The FBI arrested him October 9; OpenAI says its own AI identified him.
Published 14 October 2026 · 18:00 GMT
The device sat behind a television in one of OpenAI's executive conference rooms, plugged into a USB port, wrapped in toilet paper and, according to federal investigators, holding encrypted audio it was built to hide. On October 9, the FBI arrested the man it says put it there: 35-year-old Saudi engineer Abdulrahman Alfozan.
Federal authorities unveiled the case on Tuesday, October 13, when the Department of Justice announced that Alfozan had been arrested the previous Friday and charged with using or endeavoring to use an electronic device to intercept oral communications. He made his first appearance in federal court in San Francisco on Tuesday and remains in federal custody. If convicted, he faces up to five years in prison, a $250,000 fine and as many as three years of supervised release. A detention hearing is scheduled for Monday, October 19. Every charge at this stage is an allegation; he is presumed innocent.
OpenAI used its own AI to find the man it says bugged its boardroom.
The device, described in court filings as a surreptitious audio recording device, was thumb-sized and black, with the words "Seeed studio" printed on its casing. It carried a microphone and could be connected to a camera, drew power from the television's USB port, and was hidden behind the screen, wrapped in toilet paper. According to the Justice Department, it stored its recordings with advanced encryption and carried a data-transfer mechanism designed to evade detection by security monitoring.
OpenAI's security team found it on September 30 during a routine sweep of an executive conference room in one of the company's San Francisco buildings, and determined it had been in place for roughly two days — installed on or around September 28. That room, prosecutors say, frequently hosted discussions of highly confidential technical, commercial and strategic matters, including proprietary research, product plans, government engagements and commercial relationships.
The complaint builds its case on surveillance footage and search history. Security cameras allegedly showed Alfozan near the executive conference room on multiple days before the discovery, even though he was not authorized to enter or attend meetings there. After the sweep, investigators placed a camera in the room; it allegedly captured him entering on October 5 and looking behind the television where the device had sat. That same day, an unattended bag containing his wallet, identification and a second, similar recording device was found; he later reported the bag missing and retrieved it from the company's lost-and-found. OpenAI said it has evidence he intended to place additional listening devices in sensitive areas of its offices.
Alfozan, 35, is a Saudi national who has been a United States permanent resident since September 2020. He studied computer science and engineering at the Massachusetts Institute of Technology and earned a master's degree in data science from the University of California, Berkeley. He spent five years at Meta as a senior systems engineer, worked as a founding engineer at the logistics startup Nash, and was listed as a venture partner at Raed Ventures, a Riyadh-based investment firm. His LinkedIn profile lists the title "Engineering Lead"; OpenAI says he joined its product team as an engineer in July 2025.
The word "Saudi" in the headlines refers to his nationality, not his employer. Last month, the Dubai-based media platform Smashi Business featured him in a YouTube interview as OpenAI's only Saudi employee, according to The Times. The Financial Times has reported that he cultivated relationships with figures in Saudi Arabia's technology sector and helped connect Saudi technology representatives with Silicon Valley. None of the public filings connects those contacts to the alleged recording, and prosecutors have named no foreign government, no foreign recipient and no motive.
FBI Director Kash Patel said the bureau views the protection of American superintelligence innovation as "one of our highest counterintelligence priorities," and described the alleged scheme as an attempt to secretly capture sensitive communications about the company's superintelligence technology. The national-security framing is explicit. But the charge itself is narrower: illegal interception of oral communications. It does not allege theft of trade secrets, and it does not accuse Alfozan of acting for a foreign power. His lawyer said the recording charge is "not stealing trade secrets or working as a foreign agent," adding that no evidence suggesting otherwise had been shared with the defense. OpenAI likewise said it has no indication that model weights, source code or customer data were taken.
OpenAI's own account stresses what was not taken. In a statement posted on X, the company said it discovered "a sophisticated recording device" hidden inside one of its offices, used its own models to understand the device and identify the employee who had planted it, and then referred the matter to the FBI. "At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data," it said, adding that the employee was terminated and that it takes protecting its technology, data and people seriously "as a matter of national importance." The statement places the incident inside a security program it says is built for threats "from insider risks to cybercriminals to nation states" — a theme that runs through our coverage of the recent resignation inside OpenAI's safety ranks.
Then there is the detail OpenAI chose to advertise: it says it used its own AI models to analyze the device, trace its origins and identify the culprit within days — the industry's frontier technology turned inward, as a corporate detective. The claim arrives the same week the industry has been arguing over warnings about AI systems that can replicate themselves. For a company selling AI as an enterprise tool, the episode doubles as a demonstration: the models do not only answer questions; they police the building.
The supporting evidence in the complaint reads like a tradecraft manual assembled from office metadata. Prosecutors allege that for about a month Alfozan used an internal productivity tool to look up building addresses, meeting rooms, calendars, executive locations and external partner meetings. They also say a chatbot account associated with his wife — sharing identifiers with his own account — submitted prompts about building and concealing equipment: a continuous audio recorder, Wi-Fi connectivity, live listening and encrypted storage. His lawyer has not disputed the charge in detail, but called the allegations "wholly out of character" for his client and promised a relentless examination of the government's evidence.
The arrest lands inside a month of escalating counterintelligence drama around American technology. On October 4, the FBI arrested Wanying Zhang, a 34-year-old real estate agent from Irvine, at Los Angeles International Airport as she tried to board a flight to Shanghai, charging her with acting as an unregistered agent of China after she allegedly surveilled the Seattle-area home of a relative of Taiwan's president and sent photographs, video and license-plate information to Chinese officials (read the full story). Federal prosecutors are calling foreign espionage against American technology "extremely pervasive."
The pattern is wider than one month. In 2022 a former Twitter employee, Ahmad Abouammo, was convicted of spying for Saudi officials seeking the identities of critics on the platform. Earlier this year a former Google employee was indicted over alleged theft of trade secrets tied to the company's Pixel phones. In September, Britain's MI5 warned of Chinese state-linked sponsorship of research by more than 100 UK academics in AI and cybersecurity. And the hardware race feeding all of it — the chips, the data centers, the foundries — is tracked in our coverage of the global AI chip earnings day. When the secrets are this valuable, the United States now treats the meeting room as a front line.
The immediate calendar is short: a detention hearing on Monday, October 19, and an FBI investigation that officials say is continuing, including forensic analysis of the encrypted device and whatever it recorded. Prosecutors may add charges as that analysis lands; the question everyone in the industry is asking is whether the recordings were ever retrieved, and by whom. Washington has given no public answer — and until it does, the case sits in an uncomfortable gap: framed as a national-security threat, charged as a wiretapping offense, with the motive still blank.
What to watch: whether the forensic work produces additional charges or additional devices; how frontier labs rewrite physical-security and access rules for employees who already hold legitimate badges; and whether lawmakers treat this as the case that forces new rules on insider threats at AI labs.
In Washington's reading, the case confirms that frontier AI has become critical infrastructure: the FBI treats the protection of superintelligence research as a top counterintelligence priority, and the interception charge shows prosecutors reaching for every available statute while the forensics continue.
The corporate lesson is the threat inside the badge. OpenAI's most sensitive rooms were penetrated not by hackers abroad but by an employee with legitimate access — which is why the company's emphasis on insider-threat programs, and its public claim to have used its own models to catch him, resonates with an industry already rewriting physical-security rules.
From Beijing's vantage point, the case will be read as further evidence that Washington is securitizing the AI race: a single unproven interception charge is presented in the language of counterintelligence and national security, even though no foreign government has been named or charged.
Chinese commentary is also likely to note the contrast with the Wanying Zhang case — where a foreign power was explicitly alleged — and to treat Alfozan's Saudi nationality as a reminder that suspicion now attaches to talent pipelines themselves, not only to adversaries, pending proof.
Across the Global South, the arrest is a story about talent: a Saudi engineer educated at MIT and Berkeley, embedded at the center of the world's most watched AI company, now at the center of a federal case — a reminder of how much of Silicon Valley's frontier work runs on foreign-born expertise.
It also lands as Gulf states pour billions into their own AI ambitions, from sovereign data centers to stakes in frontier labs. Until prosecutors name a motive, the question of whether this was personal misconduct or something larger will shape how the region's AI push is received in Washington.
Abdulrahman Alfozan, 35, a Saudi national and former OpenAI product-team engineer, was arrested on October 9 and charged under federal law with using or attempting to use a device to intercept oral communications. He appeared in federal court in San Francisco on October 13 and remains in custody ahead of a detention hearing on October 19. He is presumed innocent unless proven guilty.
It was a thumb-sized black device printed with the words "Seeed studio," fitted with a microphone and able to connect to a camera, according to the Justice Department. OpenAI security found it on September 30 plugged into a USB port behind a television in an executive conference room, wrapped in toilet paper, its recordings encrypted. Prosecutors believe it was planted around September 28. A second, similar device was found in Alfozan's unattended bag on October 5.
No. The charge concerns the alleged interception of oral communications, not trade-secret theft or acting as a foreign agent. Neither the Justice Department nor OpenAI has alleged any link to the Saudi government or any foreign power, and no motive or intended recipient has been named. "Saudi" refers to Alfozan's nationality: he is a Saudi national and a United States permanent resident since 2020, whose lawyer says the evidence shared so far supports no foreign-agent theory.
It lands in the middle of a global race for AI advantage, where the FBI says protecting American superintelligence research is a top counterintelligence priority. It follows other recent cases: the October 4 arrest of Wanying Zhang, accused of acting as an unregistered agent of China, and earlier convictions and indictments involving technology theft. The episode also shows insider threats now rank alongside hackers and nation states in how frontier labs design security.
