In the quiet corridors of one of the world's most consequential technology companies, a hidden microphone wrapped in toilet paper became a symbol of the anxieties that accompany the race to build artificial intelligence. Abdulrahman Alfozan, a 35-year-old Saudi national employed as an engineer at OpenAI, was arrested in early October 2026 after federal investigators determined he had secretly planted a recording device inside a San Francisco conference room — and was found carrying a second one before his arrest. The case, still unfolding, asks a question that no algorithm can yet answer: who,
OpenAI engineer arrested for planting recording device in conference room
A recording device wrapped in toilet paper, plugged into a television
So he just walked into a conference room and plugged a device into a TV? How does that even work?
The device was wrapped in toilet paper and connected to a USB port on the back of the television. It's a fairly simple concealment method, but it worked because nobody was looking for it. The company only found it because they were doing some kind of sweep or maintenance.
Actually, the source doesn't say how they found it. We know they found it on September 30, and we know Alfozan had planted it days before that, but the complaint doesn't explain the discovery itself. That's a gap.
Fair point. But the surveillance footage shows he was in that room multiple times without permission?
Yes—about six times in September alone. And after they installed a camera, it caught him going back to check on the device behind the television. That's what really tied it together for investigators.
Though we should note: the complaint says the original conference room didn't have a camera when the device was installed. So he had a window where he could move freely. The camera came later.
What was he trying to capture? Sensitive conversations about their AI models?
That's the assumption, given that he planted it in a conference room. But the complaint doesn't actually say what conversations he was targeting or what he planned to do with the recordings. That's still unknown.
And critically, OpenAI says they found no evidence he actually stole any model weights, source code, or customer data. So either he didn't succeed, or he was still in the early stages.
What about the second device? The one in the bag?
He left it behind in a different room before his arrest. It was in a bag with his wallet and ID. The complaint doesn't explain what he was planning to do with it or where he intended to place it.
And that's another open question: was he working alone, or was someone else involved? The complaint doesn't address that. The forward look mentions potential foreign intelligence connections, but that's speculation at this point.
Le Pouls
- A recording device disguised inside toilet paper and plugged into a conference room television was discovered by OpenAI workers on September 30, triggering an immediate internal investigation and FBI referral.
- Surveillance footage revealed the engineer had entered a restricted conference room roughly six times without authorization — and returned to check behind the television even after a new security camera had been installed.
- A second recording device was found in an unattended bag Alfozan left elsewhere in the building before his arrest, suggesting a broader, premeditated surveillance effort whose full scope remains unknown.
- OpenAI confirmed that no model weights, source code, or customer data were compromised, and the company used its own AI tools to trace the device's origin and timeline.
- Alfozan appeared in federal court without legal representation, facing charges of illegally intercepting oral communications, while the FBI investigation into his motives and any possible foreign intelligence connections continues.
In the quiet corridors of one of the world's most consequential technology companies, a hidden microphone wrapped in toilet paper became a symbol of the anxieties that accompany the race to build artificial intelligence. Abdulrahman Alfozan, a 35-year-old Saudi national employed as an engineer at OpenAI, was arrested in early October 2026 after federal investigators determined he had secretly planted a recording device inside a San Francisco conference room — and was found carrying a second one before his arrest. The case, still unfolding, asks a question that no algorithm can yet answer: who, in the end, is watching the watchers?
On September 30, workers at OpenAI found something that did not belong: a recording device, wrapped in toilet paper and plugged into the back of a conference room television. Within days, the discovery had produced a federal arrest and a wiretapping charge.
The man charged was Abdulrahman Alfozan, a 35-year-old Saudi national who had been working as an engineer at the company. Arrested a week after the device was found, he made his first court appearance on Tuesday. OpenAI moved swiftly — using its own AI models to analyze the device and determine when it had been planted, then reporting the matter to the FBI and terminating Alfozan's employment. The company confirmed that no sensitive data, source code, or model weights had been accessed or stolen.
But the criminal complaint painted a more unsettling picture than a single rogue device. Closed-circuit footage showed Alfozan entering the restricted conference room on roughly six occasions in September without authorization. After the first device was found and a camera installed, he returned anyway — and was captured on film checking the spot behind the television where the original device had been hidden. Records also showed he had used a company tool to search for building addresses, meeting rooms, and internal calendars, suggesting he was methodically mapping the physical space.
Before his arrest, Alfozan left an unattended bag in another part of the building. Inside were his wallet, his identification, and a second recording device. What he intended to capture — and for whom — remains unanswered. The FBI investigation is ongoing, and questions about motive and any possible foreign intelligence connections have yet to be resolved. Alfozan appeared in court without a lawyer to speak on his behalf. The case leaves the broader AI industry with a harder question: how much of what happens inside these buildings can truly be kept secret?
On September 30, workers at OpenAI discovered something that should not have been there: a recording device, carefully wrapped in toilet paper and plugged into the back of a television set inside one of the company's conference rooms. The discovery set off a chain of events that would lead, within days, to a federal arrest and charges of illegal wiretapping.
The employee responsible was Abdulrahman Alfozan, a 35-year-old Saudi national working as an engineer at the company. He was arrested on Friday—a week after the device was found—and made his first court appearance on Tuesday, facing a charge of using an electronic device to intercept oral communications. The criminal complaint, filed in federal court in San Francisco, laid out the particulars of what investigators believe he had done and how they caught him.
OpenAI's own investigation moved quickly. The company used its AI models to analyze the device and trace its origin, determining that Alfozan had planted it days before it was discovered. Once they understood what had happened, OpenAI reported the matter to the FBI and terminated Alfozan's employment. In a statement, the company said it had found "a sophisticated recording device that was hidden inside one of our offices" and praised the speed with which federal agents and prosecutors responded. The company also made clear what it had not found: no evidence that Alfozan had accessed or stolen model weights, source code, or customer data.
But the complaint reveals a pattern of suspicious behavior that extended beyond the single device. Closed-circuit television footage showed Alfozan near the conference room on roughly half a dozen occasions in September, despite having no authorization to be in that space. After the first recording device was discovered and a camera was installed in the room, that new camera captured him entering the space and checking behind the television where the original device had been hidden. The complaint also notes that Alfozan had used a company productivity tool to search for information about OpenAI's building addresses, meeting rooms, room capacities, and calendars—queries that suggested he was mapping the physical space.
There was more. Before his arrest, Alfozan left an unattended bag in a different part of the building. Inside were his wallet, his identification, and a second recording device. The complaint does not explain what he intended to do with it or what he hoped to capture with either device. It also does not detail what conversations or information he was trying to obtain, or whether his actions were connected to any foreign intelligence service or other outside party. Those questions remain open as the FBI's investigation continues.
Alfozan appeared in court without a lawyer who could speak on his behalf. OpenAI said it would continue to support the FBI's work. The case underscores the security challenges facing major artificial intelligence companies as they develop increasingly valuable technology—and raises hard questions about how an employee with access to sensitive spaces was able to operate as he apparently did.
Citations marquantes
We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier.— OpenAI spokesperson
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data.— OpenAI statement