An engineer at the artificial intelligence company OpenAI has been arrested and charged with attempting to bug confidential meetings by concealing recording devices, according to federal prosecutors.
Abdulrahman Alfozan, 35, a Saudi national who has worked for OpenAI since July 2025, was taken into federal custody on Friday. Prosecutors allege he attempted to surreptitiously record high-level discussions involving the company's proprietary research, technical capabilities, and government relations.
According to the complaint, OpenAI discovered a "surreptitious audio recording device" hidden inside a USB port on the back of a television in a conference room, wrapped in toilet paper in an effort to disguise it. Footage showed Alfozan near the conference room on multiple occasions, despite not having authorization to access it or attend the meetings, the complaint stated.
Authorities also reported that Alfozan used chatbots to plan the alleged eavesdropping, inquiring about OpenAI's building addresses, meeting room details, capacities, and calendars. He allegedly sought information about the locations of "leadership and exec." Additionally, another chatbot account linked to his wife reportedly contained queries on how to install and conceal audio recording devices.
On October 5, surveillance footage allegedly captured Alfozan entering the conference room, possibly to inspect the recording device. Around the same time, OpenAI security personnel discovered a similar recording device in Alfozan's unattended bag, the Department of Justice said.
If convicted, Alfozan faces a maximum sentence of five years in prison and a $250,000 fine, followed by three years of supervised release.
The incident occurs amid a global race for AI dominance, with the U.S. striving to protect its technological advancements from adversaries. FBI Director Kash Patel stated that protecting U.S. "Super Intelligence innovation is essential to safeguarding our nation’s economic strength and national security," and that the FBI prioritizes this mission as one of its highest counterintelligence efforts.
An OpenAI spokesperson confirmed that there was no indication Alfozan had accessed or exfiltrated model weights, source code, or customer data. The company stated that its security program is designed to counter various threats, including insider risks, cybercriminals, and nation-states, and that this case demonstrates how AI can aid in addressing sophisticated security challenges. Alfozan has been terminated from his position at OpenAI.