“Apple is getting this wrong”: OpenAI answers Apple’s trade secrets lawsuit with receipts

원문 링크: WordPress 원문

AI NOTES · EN ENGLISH EDITION

"Apple is getting this wrong": OpenAI answers Apple's trade secrets lawsuit with receipts

On July 10, 2026, Apple sued OpenAI, accusing former Apple employees of stealing trade secrets for OpenAI's benefit. On August 3, OpenAI answered with a blog post titled "Apple is getting this wrong" — attaching email exchanges and iMessage screenshots directly in the body rather than issuing a statement. A day later Apple moved for a preliminary injunction, with a hearing set for October 1. This is how a lawsuit became an evidence war in twenty-four days.

KO · 한국어 / EN · English BILINGUAL PAIR

On July 10, 2026, Apple sued OpenAI, accusing former Apple employees of stealing trade secrets “for the benefit of OpenAI.” Twenty-four days later, on the night of August 3 (US time), OpenAI published a long blog post titled “Apple is getting this wrong.” It was not a statement of a few careful sentences. The post embedded email exchanges and iMessage screenshots directly in the body — turning a courtroom dispute into a public evidence war.

The format of the rebuttal was the message

OpenAI’s post appeared overnight, signed simply by OpenAI. As 9to5Mac put it: “It’s not a legal response, but it’s more than OpenAI has said (or shown) in the three and a half weeks since Apple’s initial lawsuit.”

The first paragraph drew its line immediately. “Apple is one of the greatest companies of all time, and built a reputation for obsessing over the smallest details,” the post began. “This careless, aggressive and oddly personal lawsuit sadly doesn’t live up to that reputation.” Those three adjectives — careless, aggressive, oddly personal — have since become the standard shorthand for the case.

The format itself carried meaning. It is rare for a litigant to publish emails and messenger captures on its own corporate blog. It blurs the boundary between public persuasion and legal defense, and OpenAI chose to blur it deliberately.

The shape of Apple’s complaint: interviews as intelligence channels

Apple filed in the U.S. District Court for the Northern District of California. The defendants are former Apple employees Chang Liu and Tang Tan, plus OpenAI and io Products. Tang Tan was Apple’s VP of product design, leading iPhone and Apple Watch design, before leaving in February 2024 to work with Jony Ive. Chang Liu spent eight years at Apple as a senior system electrical engineer before joining OpenAI in January 2026.

The complaint’s most recurring scene is the job interview. Apple alleges that Tang Tan used insider knowledge of confidential projects to grill candidates, once invoking an Apple internal project codename to ask, “What’s the plan?” for an unannounced product. The complaint says he directed candidates still employed at Apple to bring “Actual parts” to interviews for “show and tell” sessions. One candidate reportedly remarked that he “didn’t even know we could take those from the office.”

According to Apple, OpenAI instructed candidates to bring “CAD/design artifacts” and “prototypes” and to disclose “subsystem and component selection,” the “tools or methodologies” used for system integration, and “Vendor selection and communication/collaboration with vendors.” In one account, a candidate began screenshotting and downloading files on a “highly confidential Apple project” hours before interviewing with Tang Tan, who then solicited more information about that same project once the interview started. Apple calls this an “established pattern.”

The complaint also alleges that Tang Tan possessed and distributed an internal Apple “Need to Know” document — including Apple’s departure security protocols — to new OpenAI hires before they gave notice to Apple. Apple’s investigation, it says, found “a pattern by employees who depart for OpenAI of taking steps to evade the security processes intended to protect Apple’s confidential information.”

Chang Liu’s case: a security bug, a thousand pages, an unreturned laptop

The allegations against Chang Liu are equally granular. Apple claims he exploited a security bug to download confidential engineering files after leaving the company — and that rather than report the exploit, he joked about it in messages (“LOL,” “so funny”). The downloaded material, per the complaint, was a compilation of technical files running over a thousand pages, including detailed manufacturing documents for the complex circuit boards used in Apple hardware. Liu also allegedly failed to return an Apple-issued laptop.

Apple further alleges that Liu coached another Apple employee he was recruiting to OpenAI on which confidential materials to study before her own interview.

Two supply-chain claims round out the picture. Apple says OpenAI had a trusted Apple partner perform Apple’s proprietary metal-finishing technique, misleading the partner into believing Apple had authorized it. And it says OpenAI approached a second longtime Apple supplier working on power and battery manufacturing, using insider terminology to ask “targeted questions” about specific Apple components.

The complaint describes all of this as “the tip of the iceberg,” noting Apple cannot see what happens behind closed doors at OpenAI. It also states that more than 400 former Apple employees now work at OpenAI.

One misdirected email on February 23, 2026. According to OpenAI, no specific allegations arrived for five months after that.

One email in February: the wrong recipient and five months of silence

Half of OpenAI’s rebuttal is not about the substance of the theft allegations but about procedure. Apple’s complaint claimed it raised concerns with OpenAI in February and never received a response. OpenAI’s version is the mirror image.

According to OpenAI, Apple’s outside lawyers emailed the wrong person after confusing two Asian last names — a mistake Apple admitted only after OpenAI pointed it out. OpenAI also says Apple claimed a discussion with OpenAI’s General Counsel that “they now concede never happened,” and that at the time Apple told OpenAI it was “resolving any issues.” Then nothing for five months, until the lawsuit.

OpenAI attached email correspondence to make the point. The effect was to convert a claim-versus-claim standoff into a document-versus-document one. 9to5Mac had already reported, on July 15, how the email mistake derailed pre-lawsuit talks between the two companies.

OpenAI’s defense: “we do not have, nor want, any of their trade secrets”

On Chang Liu, OpenAI introduced a new frame. It says Apple employees reached out to Liu themselves and asked for his help locating the information in question — messages it attached as evidence. OpenAI argues that Apple is now trying to shift the blame to “residual access” without disclosing that this “is a common issue with Apple which is caused by them failing to properly manage system access when people leave.” In practice, the post claims, former employees trying to do the right thing when departing still find themselves with access to Apple files they never wanted and may not even have been aware of.

On Tang Tan, the defense was brief and categorical: “Tang has always been clear with the team that we do not want, and must not use, any confidential information from other companies.” OpenAI noted that Tan served Apple for more than 24 years and was “widely known as one of the most innovative leaders at the company.”

The conclusion compresses into two claims: Apple’s request for a preliminary injunction is “both based on false information and completely unnecessary,” and OpenAI does “not have, nor want, any of their trade secrets.”

iMessage records published by OpenAI (January 22, 2026). OpenAI argues Chang Liu’s access came in response to requests from Apple employees.

The same evidence, two narratives

The most interesting feature of this dispute is that both sides are presenting the same categories of evidence in opposite directions.

Consider Liu’s post-departure access. For Apple, it is evidence of exfiltration. For OpenAI, it is evidence of Apple’s own access-management failure and of Liu answering Apple employees’ requests for help. The message screenshots OpenAI published exist to serve the second reading.

Consider the February contact. For Apple, it proves “we warned them and they ignored us.” For OpenAI, it proves “the warning went to the wrong person, and we never heard the specific allegations.”

That symmetry is why this case had to become an evidence war. When the court reaches the injunction motion, it will weigh documents, not narratives.

Preliminary injunction and expedited discovery: the road to October 1

The day after OpenAI’s post, on August 4, Apple moved for a preliminary injunction. Apple says it had sent OpenAI a letter after filing, offering to forgo the motion if OpenAI agreed to five conditions. According to Apple, OpenAI agreed to the first three, and the parties could not reach agreement on the last two.

Apple now asks the court for four things: immediately stop OpenAI from using or obtaining any Apple trade secrets; preserve all evidence; allow Apple to conduct forensic inspections of OpenAI’s devices and accounts; and recover any Apple confidential information still in OpenAI’s possession. “The harm is happening now,” Apple writes. “Every day that passes without an injunction allows OpenAI to embed their knowledge of Apple’s stolen information into its hardware development efforts, further damaging Apple and making it increasingly more difficult to unwind the harm.”

Apple also filed a separate motion for expedited discovery, including early document production and depositions of key OpenAI employees and executives. The hearing on the injunction is set for October 1, 2026.

A preliminary injunction motion and a request for expedited discovery. The October 1, 2026 hearing will set the direction of this case.

The shadow of a hardware war

This is not an ordinary departing-employee dispute. It lands at the moment OpenAI is preparing to bring its first consumer hardware to market.

OpenAI’s hardware effort is led by Jony Ive. OpenAI acquired Ive’s startup io last year in a $6.5 billion deal that included more than 50 engineers, developers, and other employees. Ive founded io together with Scott Cannon, Evans Hankey, and Tang Tan. Hankey led Apple’s design team for several years after Ive’s departure, and Cannon also previously worked at Apple. Notably, none of the three — Ive, Hankey, Cannon — is named personally in Apple’s initial filing.

There is a second tension in the background. In May, Bloomberg reported that OpenAI was preparing legal action against Apple over how the ChatGPT-Siri partnership played out. Apple’s complaint says that agreement is not at issue here.

OpenAI’s hardware roadmap keeps sharpening. In April, Ming-Chi Kuo reported that OpenAI is developing its own smartphone, potentially launching in 2028. The Information has reported on OpenAI’s work on a HomePod-style smart speaker. Every step that roadmap takes toward reality makes the question at the center of this litigation — which knowledge came from where — heavier.

Three things to watch next

First, the October 1 injunction hearing. If the injunction is granted, parts of OpenAI’s hardware development process will operate under court supervision. If it is denied, the case settles in for a long merits fight.

Second, the scope of discovery. If early document production and depositions are allowed, the fight will turn on how much of OpenAI’s hiring process — and what departing employees brought with them — enters the record.

Third, the names not in the complaint. Whether io’s central figures — Ive, Hankey, Cannon — appear in an amended complaint or in discovery will be the best gauge of how wide this case expands.

The verdict in this fight will come from documents, not public opinion. OpenAI’s decision to publish its evidence on the company blog reflects a calculation that the side that shows its documents first controls the narrative. In the two months before October 1, both sides will keep talking in documents.

Sources

• OpenAI rebuttal, “Apple is getting this wrong”: https://openai.com/index/apple-is-getting-this-wrong/

• Apple’s complaint (PDF, via 9to5Mac): https://9to5mac.com/wp-content/uploads/sites/6/2026/07/Apple-Inc.-v.-Liu-et-al.pdf

• Apple sues OpenAI (9to5Mac, 2026-07-10): https://9to5mac.com/2026-07-10/apple-sues-openai-trade-secret-theft/

• OpenAI rebuts the allegations (9to5Mac, 2026-08-03): https://9to5mac.com/2026-08-03/openai-rebuts-apple-trade-secrets-allegations-in-new-response-and-evidence/

• Apple moves for preliminary injunction (9to5Mac, 2026-08-04): https://9to5mac.com/2026-08-04/apple-preliminary-injunction-openai/

• OpenAI’s first response (9to5Mac, 2026-07-10): https://9to5mac.com/2026-07-10/openai-responds-to-apples-trade-secret-theft-lawsuit/