Apple accuses ex-engineer of using its trade secrets to train an AI agent at OpenAI
Apple says a former senior engineer didn’t just take confidential information after leaving the iPhone maker for OpenAI. The company now alleges he used a proprietary Apple circuit schematic in his new job and trained an AI agent to perform engineering work with it, adding a new twist to an already bitter trade secret fight between two of Silicon Valley’s biggest names.
The allegations center on Chang Liu, a former senior system electrical engineer at Apple who joined OpenAI in January 2026. In a federal court filing Monday, Apple said a forensic examination of Liu’s MacBook uncovered evidence that he used a confidential Apple file containing a power-converter circuit schematic after he joined OpenAI.
“The ongoing forensic inspection of this MacBook has uncovered forensic evidence showing that this computer contains Apple trade secret information, including at least one specific ‘.asc’ file, and that this file has been used while Mr. Liu has been employed by OpenAI,” Apple said in the court filing.
According to Apple, Liu accessed the proprietary circuit information in March, months after leaving the company. Reuters reported that Apple alleges Liu then trained an AI agent using proprietary Apple information.
The details make this case unusual. Apple claims Liu ran the confidential circuit schematic through LTspice, an engineering application used to design and simulate electrical circuits. Messages cited in Apple’s filing reportedly show Liu discussing an AI “agent” that had learned to operate LTspice, review simulation results, and adjust parameters.
Apple’s concern goes beyond one engineer opening one confidential file. The company argues that feeding trade secret information into an AI system could allow that knowledge to persist and be reused, potentially making the effects harder to contain than a conventional leak.
That argument could turn this lawsuit into an early test of a much bigger question facing companies across the AI industry: What happens when proprietary engineering knowledge leaves with an employee and becomes part of an AI-assisted workflow?
A MacBook adds a new chapter to Apple’s case
The new evidence surfaced after Liu’s lawyers delivered equipment, including a MacBook, to Apple’s forensic expert on August 21, according to Apple’s August 31 filing in the U.S. District Court for the Northern District of California.
Apple’s forensic expert began reporting preliminary findings on August 26. The company told the court that the ongoing examination found Apple trade secret information on the computer, including at least one “.asc” file that Apple says was used during Liu’s employment at OpenAI.
Apple had already identified that file in its preliminary trade secret disclosures. The company said its forensic investigation remains incomplete.
The findings prompted Apple to seek permission to supplement its push for expedited discovery. Apple wants faster access to documents and testimony to determine the scope of any alleged misuse and whether further court intervention is needed. The defendants agreed not to oppose Apple’s request to submit the supplemental filing, subject to their right to respond. Their response is scheduled for September 4.
OpenAI has sharply rejected Apple’s broader claims.
In an August 3 response, OpenAI said Apple was “getting this wrong” and stated that it does not have or want Apple’s trade secrets. The ChatGPT maker said former Apple employees retained access to some company files after leaving because Apple failed to revoke access properly. OpenAI published messages showing Apple employees contacting Liu for help locating information after his departure.
Reuters reported that OpenAI has sought to dismiss the case and says its hardware team is developing “entirely new” products and technologies.
No court has ruled that Liu or OpenAI misappropriated Apple’s trade secrets. For now, the claims remain allegations, and the forensic examination is still underway.
Yet the new filing pushes the dispute into territory that extends far beyond Apple and OpenAI. Companies have spent decades worrying about employees walking out with confidential documents. AI agents introduce a new problem: proprietary knowledge may no longer need to stay inside a document to remain useful.


