Nvidia reportedly plans about €2B in new European AI deals

OpenAI says it found a sophisticated recording device hidden inside one of its offices, then used its own models to help identify the device and trace it back to a specific employee. The company says the device had been planted a few days earlier. It also says the employee was arrested and terminated, and that the FBI is now handling the case.

The company says it does not currently believe model weights, source code, or customer data were improperly accessed or taken. That part matters as much as the spy-movie setup. If the claim holds up, this was not a breach of a server room or a cloud bucket. It was a physical device inside a real office, with AI helping investigate the scene.

OpenAI is clearly treating the incident as more than a one-off security scare. In its telling, its models acted as another layer of defense against insider threats, cybercriminals, and nation-state actors. That’s a notable shift in how the company frames AI’s role: not just building products, but helping protect the company from human trickery in the first place.

The FBI investigation is still ongoing, so the full story is not settled. But the episode says something about where the industry is headed. AI is now being used to sort through the weird, messy, physical side of security too, not just code and logs. The machine age has apparently reached office espionage.