Reuters: Apple, in its new filing says a former engineer used its confidential circuit design inside OpenAI's AI workflow.

The filing says Chang Liu, (the electrical engineer who worked at Apple) downloaded the power-converter schematic after leaving Apple, ran it in LTspice, and generated simulation outputs on a Mac mini.

Apple claims his AI agent learned to run LTspice, inspect results and tune compensation parameters, cutting his stated workflow from a day to two hours.

Apple argues this raises a distinct AI risk because proprietary information may become harder to trace or remove after an agent learns from it.

The filing does not establish that OpenAI trained model weights on Apple's data; it describes an AI agent learning to operate the simulator and tune parameters.

Ofcourse, the court has not ruled that OpenAI stole or destroyed trade secrets; Apple submitted this evidence to support its request for expedited discovery.