Apple-OpenAI dispute puts hiring practices under trade secrets microscope

As competition for artificial intelligence talent intensifies, Apple's trade secrets case against OpenAI is putting renewed attention on a longstanding question in employment law: When does recruiting experienced employees cross the line into trade secret misappropriation?

Legal experts say companies are generally free to ask job candidates about their prior work experience and expertise. The legal risk arises when an employer allegedly seeks confidential information protected by trade secret law rather than an employee's general knowledge and skills.

"Indeed, if employees cannot brag about their experience and knowledge and performance to prospective employers, they will not be able to get outside offers and mobility will be suppressed," said University of San Diego law professor Orly Lobel.

Apple alleges OpenAI crossed that line as it built its hardware division by recruiting Apple employees and soliciting confidential information during the hiring process.

The lawsuit, filed Friday in the Northern District of California, alleges former Apple engineer Chang Liu accessed confidential documents before leaving the company and instructed a coworker how to do the same. Apple also claims OpenAI Chief Hardware Officer Tang Tan used internal Apple information to question prospective hires about confidential projects.

For Apple to prevail, however, University of Virginia law professor Elizabeth A. Rowe said it will have to prove far more than that OpenAI asked candidates about work they performed at Apple.

"Ultimately, it will have to do with what we call 'improper means,'" Rowe said.

According to Rowe, the inquiry focuses on both sides of the interview.

"On the one hand we have to look at such facts as: is the person being asked ... do they realize and understand that they are bound by confidentiality to certain information?" Rowe said. "On the other side of the table as well is looking ... at the intentions and motivations and knowledge of the person [or] company asking the question, and whether they also have knowledge that the information that they are seeking is covered by confidentiality and that they're essentially asking someone to breach a duty."

Stanford University law professor Mark A. Lemley said not every allegation in Apple's complaint would constitute unlawful conduct under California law.

"Some of the things Apple alleges, like the fact of hiring many employees or 'coaching' Apple employees how to leave, aren't illegal in California and may actually be part of good business practice," Lemley said. "But if Apple's claims that the employees took confidential documents with them -- and that OpenAI is using those documents -- are true, that is a problem for OpenAI."

Apple, represented by Weil, Gotshal & Manges LLP, must first establish that the information at issue qualifies as trade secrets, Rowe said.

"The first is that it owned trade secrets -- whatever its list of information is, the kinds of technologies here that it claims to be its trade secrets, that it has taken reasonable efforts to protect these secrets, that it is not publicly known, etc.," she said.

The company must then show those trade secrets were misappropriated through improper means.

Under California's Uniform Trade Secrets Act, Rowe said, liability can arise from improperly acquiring trade secrets even before they are put to use.

"Under the Uniform Trade Secrets Act, if you're merely acquiring by improper means [it's] enough for potential liability," Rowe said. "So they don't have to use it, and we have also all kinds of very broad definitions of what 'use' means."

One challenge for both sides will be distinguishing between an employee's accumulated expertise and proprietary information learned on the job.

"The problem is sometimes it is very hard to distinguish between the two, especially when you have the kinds of really high-level employees, for instance, who participated in developing the information at issue," Rowe said.

Lobel called it "inevitable" that employees carry knowledge from one company to another.

"A company may hire someone because that person understands hardware engineering, product development, suppliers, or manufacturing," Lobel said. "The legal line is crossed when the new employer seeks the former employer's actual secret materials or asks questions calculated to elicit proprietary information rather than transferable expertise."

Evidence developed during discovery is likely to determine where that line was crossed, if at all.

"An inappropriate interview question is powerful evidence of intent, but Apple generally must show more than curiosity or aggressive questioning," Lobel said. "It will want evidence that specific confidential information was disclosed, retained, transmitted, incorporated into OpenAI's work, or used to accelerate a competing hardware program. Device records, downloads, messages, interview notes, design overlaps, and the timing of development could therefore be critical."

The role of Tan may also receive particular scrutiny. Before becoming OpenAI's chief hardware officer, Tan served as Apple's vice president of product design for the iPhone and Apple Watch. He later co-founded io Products, a startup that OpenAI acquired for more than $6 billion.

"Sometimes that structure can be seen as a way to disguise the exchange of information," Rowe said. "That's another thing that I'm sure they will do a lot of discovery on."

Because Apple is seeking a preliminary injunction, Rowe said the parties could enter expedited discovery, bringing key evidence to light much earlier than in a typical trade secrets case.

Once the court rules on the injunction request, she added, the litigation could become more conducive to settlement because both sides will have an early indication of how the judge views the merits.

The stakes extend beyond the courtroom. OpenAI is widely rumored to be developing hardware products that could compete with Apple, making any delay potentially significant.

An injunction "is incredibly important to these companies," Rowe said, because it could affect "how quickly OpenAI gets a product to market," underscoring "why the injunctive power here is really tied to the stakes of the game from the companies' perspective."

The case is Apple Inc. v. Liu et al., 5:26-cv-07078 (N.D. Cal., filed July 10, 2026).