FBI Arrests OpenAI Engineer Accused of Hiding a Recording Device in an Executive Meeting Room

The FBI arrested OpenAI engineer Abdulrahman Alfozan on October 9, 2026 on a charge of hiding an audio recording device in an executive conference room at one of OpenAI's San Francisco buildings, the US Justice Department said on October 13. OpenAI staff found the device on September 30, wrapped in toilet paper and plugged into a USB port on the back of a television, according to the department's announcement.

Alfozan, 35, is a Saudi national who has been a legal permanent resident of the US since September 2020. He faces one count of using or trying to use a device to intercept oral communications, a federal wiretap offense, and he made his first court appearance on October 13. The charge is an allegation, not a conviction. The case adds to a tense stretch for AI security and policy at OpenAI, which confirmed on October 1 that it had fired three safety researchers over the handling of sensitive information.

What is the OpenAI engineer accused of doing?

Prosecutors allege that Alfozan installed a thumb-sized recording device on or about September 28, 2026 in an executive conference room he had no authority to enter, according to the FBI affidavit filed with the criminal complaint. Only executives and their assistants can reserve the room, which is used for leadership meetings, technical and product sessions, and briefings with government officials and business partners, the affidavit says.

The device was a Seeed XIAO ESP32-S3 Sense, a small development board with a microphone and microSD storage. OpenAI's own investigation found that its memory card's contents may have been encrypted, and no files could be recovered. The firmware appeared to have been compiled on September 28, which is how investigators dated the installation.

FBI Special Agent Lydia Liang swore the complaint on October 8; it calls OpenAI only "Company 1", while the Justice Department release names the company.

How did investigators tie the device to Alfozan?

The FBI's case rests on office camera footage, Alfozan's questions to a workplace productivity agent and prompts sent from an AI chatbot account linked to his spouse, according to the affidavit. There was no camera inside the room when the device went in, but cameras elsewhere in the office recorded him near the room on 10 dates between September 8 and September 29, including from 2:18 pm to 6:30 pm on September 28.

Between August 17 and September 16, Alfozan asked the productivity agent about building addresses, meeting rooms, room calendars, where "leadership and exec" were based and where external partners usually met, the affidavit says. From September 24 to 30, the spouse's account, which shared installation identifiers, IP usage and user path names with his own account, sent prompts about building a continuous audio recorder, live listening, encrypting the memory card, speech-triggered recording and a second and third unit. Prompts sent on September 27 and 28 included the found device's unique MAC address (a hardware identifier) and its custom storage header.

By October 5 a camera had been placed in the room, and it filmed Alfozan entering at about 1:20 pm and checking behind the television. The same day he left a bag in a printer room; security staff opened it and found his wallet, his ID and a second Seeed recording device. When he collected the bag on October 6, he reported missing items, including a credit card and about $500 in cash, but did not ask about the device, according to the affidavit.

What has OpenAI said about the arrest?

OpenAI said in a post on X on October 13, 2026 that it has no indication the employee accessed or took model weights, source code or customer data, and that it fired him. The company said it used its own models to understand the device and trace it to the employee, then referred the matter to the FBI.

OpenAI spokesperson Nate Evans told TechCrunch that Alfozan joined in July 2025 and worked on the product team, and the outlet reports he was fired and arrested on the same day. OpenAI also told TechCrunch the device had "a mechanism for transferring data designed to evade monitoring."

The two accounts put the weight in different places. OpenAI credits its models with identifying the employee, while the sworn affidavit builds its attribution on camera footage, agent queries and chatbot prompts and does not mention OpenAI's models.

Who is Abdulrahman Alfozan?

Alfozan is listed as a partner at Raed Ventures, a Riyadh-based venture firm focused on the Middle East and North Africa, which announced him as a venture partner in February 2025, according to CNBC. That announcement lists past roles at Meta, Microsoft and SAP and computer science studies at MIT, CNBC reports; Reuters said his LinkedIn profile also lists the University of California, Berkeley, which it could not verify (Reuters via CBC).

Neither the Justice Department release nor the affidavit says for whom any recordings were intended. TechCrunch and Reuters said they could not reach Alfozan or a lawyer for comment.

How are US officials framing the case?

The Justice Department presented the arrest as a national security matter and said the prosecution supports National Security Presidential Memorandum 11, which it describes as designed to support American superiority in super intelligence technology. Director of National Intelligence Jay Clayton, who chairs the Super Intelligence Force that President Donald Trump announced on October 4, 2026, cited the case as a reason for it.

FBI Director Kash Patel called protecting the industry "one of our highest counterintelligence priorities" in the release. The charge carries up to 5 years in prison, a $250,000 fine and 3 years of supervised release.

What we don't know yet

• For whom, if anyone, the recordings were intended, and whether others were involved.
• What, if anything, the first device recorded before it was found, since its contents could not be decrypted.
• Whether a grand jury will indict Alfozan, who represents him and how he will plead.
• How OpenAI's models helped identify the employee, beyond the company's one-line description.