Inside one of the world's most closely watched artificial intelligence companies, a hidden microphone wrapped in toilet paper became the quiet center of a federal investigation. Abdulrahman Alfozan, a 35-year-old Saudi national employed as an engineer at OpenAI, was arrested by the FBI after the company discovered he had covertly planted a recording device in a conference room — and, it emerged, carried a second one in an unattended bag. The case raises questions that the charges alone cannot answer: what was meant to be heard, and by whom.
OpenAI engineer arrested for planting listening device in conference room
A pattern of unauthorized access, caught on camera
Mark
Why would an engineer working inside OpenAI need to plant a hidden microphone? Wouldn't he already have access to meetings?
Mimi
That's the question nobody's answered yet. The complaint doesn't say what he wanted the recordings for, or whether he was acting alone. He clearly had some access to the building, but the surveillance shows him going to places he shouldn't have been—six times in one month.
Luke
Right, and that matters. He had *some* legitimate access as an employee, but the pattern suggests he was specifically trying to get into spaces where he didn't belong. The second device in the bag is interesting too—it suggests this wasn't a one-off impulse.
Mark
Did he actually record anything sensitive? Did he get what he was after?
Mimi
OpenAI says they found no evidence he stole model weights, source code, or customer data. But that's about data theft, not about whether he captured conversations on that microphone.
Luke
Exactly. The company can tell you whether their AI systems were compromised. They can't necessarily tell you what was on that recording device before they found it. Those are two different things.
Mark
So he could have recorded something and passed it on before they caught him?
Mimi
Theoretically, yes. But there's no indication of that in the complaint. The complaint is focused on the act of planting the device itself—the charge is about the interception attempt, not about what happened with any recordings.
Luke
And we don't know his motive. Corporate espionage? Selling information? Something else entirely? The complaint is silent on that, which is unusual for a case this serious.
Mark
What does it tell us that he's a Saudi national?
Mimi
The complaint identifies him that way, but it doesn't say whether that's relevant to the investigation. It's a fact, but we shouldn't read too much into it without more information.
Luke
The FBI investigation is ongoing, which means there's a lot we still don't know. This is the arrest, not the full story.
Il Polso
- A surveillance device disguised in toilet paper and plugged into a television's USB port sat undetected inside an OpenAI conference room until September 30th, when the company found it.
- Closed-circuit footage revealed Alfozan had entered the room without authorization at least six times in September, and returned after a new camera was installed — appearing to check on the hidden equipment.
- A second recording device was found in an unattended bag Alfozan left elsewhere in the building before his arrest, deepening the scope of the suspected operation.
- OpenAI deployed its own AI models to analyze the first device and trace when it was planted, then reported its findings to the FBI and terminated Alfozan's employment.
- Federal charges have been filed, but the criminal complaint leaves the most pressing questions unanswered — including Alfozan's motive and whether he was acting alone or on behalf of another party.
Inside one of the world's most closely watched artificial intelligence companies, a hidden microphone wrapped in toilet paper became the quiet center of a federal investigation. Abdulrahman Alfozan, a 35-year-old Saudi national employed as an engineer at OpenAI, was arrested by the FBI after the company discovered he had covertly planted a recording device in a conference room — and, it emerged, carried a second one in an unattended bag. The case raises questions that the charges alone cannot answer: what was meant to be heard, and by whom.
On September 30th, OpenAI discovered something unsettling inside one of its conference rooms: a sophisticated recording device, hidden in toilet paper and connected to a USB port on the back of a television. Within days, the company had identified the person responsible. By Friday, the FBI had made an arrest.
The employee was Abdulrahman Alfozan, a 35-year-old Saudi national working as an engineer at the company. He appeared in federal court in San Francisco on Tuesday, charged with using an electronic device to intercept oral communications. The criminal complaint does not explain what Alfozan intended to do with the recordings, or whether he was working with or for anyone else.
What the complaint does establish is a deliberate pattern. Surveillance footage showed Alfozan near the conference room on at least six occasions in September, despite having no legitimate reason to be there. After OpenAI installed a new camera following the discovery, that footage captured him entering the room and checking behind the television where the device had been placed. Records also showed he had used an internal tool to repeatedly search for building addresses, meeting rooms, and calendar schedules.
Before his arrest, Alfozan left an unattended bag in another part of the building. Inside were his wallet, identification, and a second recording device. OpenAI used its own AI models to analyze the first device and determine when it had been installed, then reported its findings to the FBI and terminated his employment. The company stated it had found no evidence that AI model weights, source code, or customer data had been accessed or stolen. The investigation remains ongoing.
On September 30th, OpenAI discovered a recording device hidden inside one of its conference rooms—a sophisticated piece of surveillance equipment concealed in toilet paper and wired into a USB port on the back of a television. Within days, the company had identified who planted it. By Friday, the FBI had made an arrest.
The employee was Abdulrahman Alfozan, a 35-year-old Saudi national who worked as an engineer at the company. He appeared in federal court in San Francisco on Tuesday, charged with using an electronic device to intercept oral communications. The criminal complaint filed in the case does not explain what Alfozan intended to do with recordings of the company's internal conversations, or who, if anyone, he was working with or for.
What the complaint does establish is a pattern of unauthorized access. Closed-circuit footage from inside OpenAI's offices showed Alfozan near the conference room on at least six separate occasions in September, despite having no legitimate reason to be there. After the company installed a camera in the room following the discovery of the device, that new footage captured him entering the space and checking behind the television where the recording equipment had been placed. The complaint also notes that Alfozan used an internal productivity tool to search repeatedly for information about the company's building addresses, meeting rooms, room capacities, and calendar schedules.
The investigation took another turn when, before his arrest, Alfozan left an unattended bag in a different part of the building. Inside were his wallet, identification, and a second recording device. OpenAI used its own artificial intelligence models to analyze the first device and trace its origins, determining that Alfozan had installed it days before it was discovered. The company reported the finding to the FBI and terminated Alfozan's employment.
In a statement, an OpenAI spokesperson said the company had "no indications" that Alfozan had inappropriately accessed or stolen the company's AI model weights, source code, or customer data. The FBI investigation remains ongoing. It was unclear as of Tuesday whether Alfozan had retained legal representation.
Citazioni salienti
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 spokesperson
Vuoi la storia completa?
Leggi l'originale su The Globe and Mail ↗