A former OpenAI engineer has been charged in the US with allegedly planting a covert recording device in the company’s San Francisco office, a case that lands awkwardly for an industry that sells intelligence and still relies on old-fashioned trust.
Federal prosecutors said Abdulrahman Alfozan, a Saudi national who had worked at OpenAI since July 2025, was charged with using or attempting to use a device to intercept oral communications. According to the criminal complaint described by prosecutors and the BBC, the recorder was found on 30 September, inserted into a USB port behind a television in an executive conference room and wrapped in toilet paper. Prosecutors said it had been installed two days earlier.
The useful fact here is not that the device was “sophisticated,” though OpenAI used that word. It is that prosecutors say it was encrypted and allegedly placed in a room used for executive discussions about AI products. That narrows the claimed target: not a random office prank, but corporate information.
OpenAI said it alerted the FBI after discovering the device and that it used its own AI models to help identify the employee who had allegedly planted it. Prosecutors also allege Alfozan used a productivity tool and a chatbot to research sensitive locations in the office and how to build, conceal and remotely access audio-recording hardware. That does not prove anything about AI systems beyond their obvious utility as research aides. A search box is still a search box, even when dressed for a conference stage.
Prosecutors said CCTV footage showed Alfozan returning on 5 October to inspect the device. They also allege security staff found a second, similar recorder in a bag he left elsewhere in the building. OpenAI said it has no indication source code, model weights or customer data were accessed or taken. He has been fired. The BBC reported it was unclear whether he had legal representation at his court appearance.
Alex de Valletta 
Dua Mifsud