OpenAI has fired back at Apple’s trade secrets lawsuit with a defense that flips the case on its head. Rather than arguing over what its new hires did or didn’t take, OpenAI is going after Apple’s own security housekeeping. According to TechCrunch AI, the company’s motion to dismiss, along with newly filed exhibits, argues that Apple was so loose with its systems that the information it now calls stolen may not legally count as a trade secret at all.
What OpenAI is actually arguing
Apple sued in July, accusing OpenAI of running a scheme to pull confidential hardware information out of former Apple engineers. This week, Apple pushed the court to expedite discovery, saying its internal investigation suggests more ex-employees may have taken part in or witnessed the alleged theft.
OpenAI’s response, as detailed in TechCrunch AI, sidesteps the who-did-what fight entirely. Its core claims:
- Apple let employees use personal iCloud accounts for work.
- Apple failed to revoke access after people left.
- One Apple manager stayed logged into the personal iCloud account of former engineer and defendant Chang Liu after he departed, used it to transfer files, then later asked him for help with technical questions about Apple projects.
OpenAI says Apple left those “inexplicable information-management practices” out of its original complaint. The company frames the whole situation as confusion created by Apple’s own sloppiness, not theft, and argues its new hires were simply helping former colleagues.
Why the argument matters
Yes, it can read a bit like “the door was unlocked, so it wasn’t really stealing.” But there’s real legal weight here. Under trade secret law, a company has to take reasonable steps to keep information secret for it to qualify as a protected trade secret in the first place. If Apple can’t show it locked things down, the claim gets shakier.
OpenAI also hammers a second point: Apple never named the specific secrets. The filing says Apple describes them as “generic categories of the product-development process,” like component manufacturing, product testing, vendor relationships, and distribution channels. That’s a familiar defense move. Vague trade secret claims are harder to win because the plaintiff has to identify what was taken with enough precision to defend it.
The bigger fight underneath
What stands out here is the motive OpenAI assigns to Apple. The filing argues the lawsuit is really about slowing a competitor down in AI-powered hardware, not protecting Apple’s roadmap.
“OpenAI has no use, need or desire for Apple’s trade secrets,” the motion reads. “OpenAI is building something entirely new and different from anything at Apple.” It goes further, framing the suit as damage control for a talent problem: “Apple should not be permitted to use a baseless and pretextual lawsuit to make up for its shortcomings in the market for talent and retaining its employees, and its failures to integrate AI into its products.”
That’s a pointed jab, and it lands in the middle of a very real trend. OpenAI, along with Meta and others, has been pulling senior engineers out of Apple as the AI hardware race heats up. Apple’s own AI struggles have been well documented. So this case sits at the intersection of two things every AI company is fighting over right now: talent and hardware.
What to watch next
This is a motion to dismiss, not a ruling, so nothing is settled. A few things worth tracking:
- Whether the judge buys the “not really secret” framing. If Apple’s offboarding really was that leaky, it weakens the foundation of the case.
- Apple’s push for expedited discovery. More named employees could widen the fight fast.
- The precedent for talent moves. How this shakes out will shape how aggressively AI firms can recruit from hardware giants, and how those giants can fight back in court.
For anyone building or hiring in AI, the practical lesson is blunt: your offboarding hygiene is now a legal asset. Revoked access, no personal accounts for work, and clear records aren’t just IT chores. They’re what a trade secret claim stands or falls on. More detail is available in the original TechCrunch AI report.