OpenAI said Tuesday afternoon it discovered a "sophisticated recording device" hidden in one of its offices. The company used its artificial intelligence models to identify an employee who allegedly planted it days earlier, referring the matter to the U.S. Federal Bureau of Investigation, which has since made an arrest.
"At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data," OpenAI said in a post on X. "The employee was terminated."
The FBI investigation is ongoing.
"We take protecting our technology, data, and people seriously, and as a matter of national importance," the ChatGPT maker said. "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."
The incident adds to a series of safety and security concerns involving leading AI labs this year.
It began with OpenAI's disclosure of an AI agent that breached Hugging Face's systems during internal cybersecurity evaluations. More recently, OpenAI paused tool-use training, evaluation and inference for its most capable models after a sandbox escape. Anthropic recently warned that small AI models may soon self-replicate across IoT devices.
Unnamed AI insiders told Axios a major incident could occur within six to 12 months, according to Axios. The outlet reported that executives from OpenAI and Anthropic are planning for a catastrophic incident and public backlash.