A recording device wrapped in toilet paper sat plugged into a USB port behind a television in an OpenAI executive conference room for two days before staff discovered it.
US authorities have opened an investigation. A 35-year-old engineer is under arrest, accused of planting the device in OpenAI’s San Francisco office. The engineer had worked at the company since July 2025.
The room wasn’t an ordinary meeting space. The complaint says it was where confidential research, future products and business partnerships were discussed, and that is why investigators are treating the case so seriously.
Tucked away behind a television
According to the complaint, employees found the device on Sept. 30. OpenAI believes it had been put there two days before that. It was wrapped in toilet paper and connected to a USB port at the back of a television.
Prosecutors say security camera footage shows the employee looking behind the TV. They also say a second device, similar to the first, turned up in the engineer’s bag.
The complaint also points to chatbot activity. It cites queries sent from a chatbot account tied to the engineer’s wife. Those queries asked how to build recording devices, how to hide them and how to access them remotely.
OpenAI credits its own models
OpenAI’s statement opens with one claim in particular: the company says its own AI helped it work out what the hardware was and connect it to a specific person.
“We discovered a sophisticated recording device that was hidden inside one of our offices. 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 said in a statement.
“We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work,” the company said.
An AI company saying its AI caught an insider is a neat narrative. The statement doesn’t say what the models found that the camera footage and the device in the bag hadn’t already revealed.
What OpenAI says was not taken
For customers, the bigger concern is whether anything actually got out. OpenAI spoke to that point directly.
“At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated,” the company said.
The opening words matter. “At this time” covers only what OpenAI has found so far. It makes no promise about what the investigation might uncover later.
OpenAI also presented the incident as a national security issue and as proof that its wider security program is working.
“We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats,” OpenAI said.
The court has not ruled
OpenAI fired the engineer and brought in the FBI. No court has ruled on any of the charges so far. The details about the footage, the second device and the chatbot queries all come from the complaint and from prosecutors.
Right now the public record contains a roll of toilet paper behind a conference room TV, a matching device in a bag and a set of chatbot questions about concealing recorders. It will be up to prosecutors to prove in court what that evidence means.