The sensitive data reportedly came to light on August 21 after OpenAI surrendered the hardware as part of court-ordered discovery proceedings. Apple alleges that in March 2026, Liu utilized this internal information to train an artificial intelligence agent. This specific accusation serves as the basis for Apple’s current request to expedite the discovery process in their broader lawsuit.

OpenAI has consistently denied the allegations, characterizing its development process as the creation of entirely new products rather than the misappropriation of intellectual property. The startup has already moved to dismiss the lawsuit, arguing that the claims lack merit. As the legal battle unfolds in U.S. federal court, the focus remains on whether internal Apple documentation directly influenced the architecture of OpenAI’s current systems.