The New OpenAI Class Action Lawsuit: This Time It’s Not About Data

For the last few years, most of the major AI legal battles have focused on data: what was collected, whether it was copyrighted, and whether AI companies had permission to use it. OpenAI is involved in a lawsuit with The New York Times over copyrighted content used in training data for its AI models. However, OpenAI’s latest legal woes aren’t about generative AI tools and scraped websites. Now it is being accused of something more tangible: stealing. Apple filed a lawsuit against the AI giant, alleging trade secret theft and claiming that it used former employees and underhanded recruitment tactics to do so.

KEY TAKEAWAYS

• The Apple lawsuit alleges that OpenAI stole trade secrets and other protected company information.
• Apple claims OpenAI benefited from former employees who retained or accessed confidential materials after leaving the company.
• The New York Times case focuses on whether OpenAI used copyrighted journalism without permission to train AI products.
• OpenAI is likely to argue that former employees can use general expertise, and that AI training is protected by fair use.
• The next stage will depend on evidence, including internal communications, training data records, recruitment practices, and output logs.

The Latest OpenAI Class Action Lawsuit: What Happened?

The realm of AI: so full of innovation, new technology, and, ironically, stolen information and copyright infringement lawsuits. The star of the most recent round of legal action is OpenAI, facing allegations of trade secret theft from Apple. Yes, that Apple, maker of iPhones.

NOTE:

The word “realm” was used on purpose and typed by a real person.

On 10 July 2026, Apple filed a lawsuit against OpenAI in the U.S. District Court for the Northern District of California. The complaint claims OpenAI used former Apple employees, questionable recruitment tactics, and confidential product information for its foray into the hardware market.

That makes this different from the usual AI “you can’t use my data” squabbles like the Reddit vs. Perplexity case a while back.

Tech companies hire from each other all the time. Developers, engineers, designers, and executives move to greener pastures, taking their skills and experience with them.

What they can’t take are internal documents, unreleased product details, design files, technical specifications, prototypes – you get the idea.

Apple says that’s exactly what happened: OpenAI didn’t just hire people who knew how to build hardware; it also tried to use them to get its hand into Apple’s cookie jar.

The makers of ChatGPT are also currently in a two-year legal battle with The New York Times over the company allegedly using copyrighted journalistic content in its AI model training data and reproducing it in outputs. We’ll get to that later and how the two cases are related.

Usually, where there’s allegations, there’s fire. But that’s a whole other conversation.

The Allegations

Apple’s complaint contains serious allegations of blatant trade secret theft, granted with some pretty ballsy methods.

Fair warning: the word ‘allegedly’ is going to be used a lot from here on.

We’ll start with the ‘show and tell sessions.’ That might sound innocent enough, but if you believe Apple’s version of events, there was nothing innocent about it.

Apple claims that OpenAI asked job candidates to bring ‘actual parts’, CAD files, prototypes, and design artifacts from Apple to interviews. To which one candidate apparently said: “Didn’t even know we could take those from the office.”

And these are the people working at the forefront of technology…

Apple also alleges that OpenAI coached departing employees on how to get around internal security procedures.

They claim this included how to avoid the “dreaded walkout”- being immediately escorted off the premises and cut off from internal networks – by hiding who they were going to work for.

This apparently let them work their 2 weeks’ notice with full access, presumably to gather more intel. OpenAI was also accused of telling them not to sign anything during exit interviews and to let them know “ASAP.”

Again, these are allegations; nothing has been proven as fact. If it does turn out to be true, OpenAI’s recruitment team has some serious explaining to do.

Here’s the thing: if a company simply hires someone from a competitor, fair enough. If it allegedly gives that person advice on how to avoid security before leaving the building, that’s something entirely different.

Based on the above, it seems former Apple staff weren’t just being asked about their skills, experience, or where they see themselves in the next 47 years during interviews. They were allegedly handing over details on products and systems that hadn’t been made public.

Of course, OpenAI has denied everything, posting a statement on X saying:

“We have no interest in other companies’ trade secrets. We remain focused on building innovative technology that empowers people everywhere.”

As far as denials go… meh.

Who Is Involved: Meet the Cast

Apart from the companies themselves, a few key players are named in Apple’s lawsuit. And there’s definitely a pattern.

Starting us off is Tang Tan, OpenAI’s Chief Hardware Officer. Before joining OpenAI, he spent 24 years at Apple, most recently as VP of product design for the iPhone and Apple Watch.

Is it a coincidence that someone so close to the design, engineering, and manufacturing processes Apple protects so fiercely is named in the case, or just sour grapes?

Apple alleges that Tan used confidential Apple project code names during OpenAI recruitment, and asked candidates about unannounced products. He was apparently also one of the people accused of asking candidates to bring components to interviews and of coaching them on how to avoid security.

Chang Liu is another former employee named in the complaint. Liu worked at Apple for 8 years as a senior systems electrical engineer before joining OpenAI in 2026. Apple alleges he failed to return a company-issued laptop and used it to download confidential technical documents.

Apple says Liu texted Yu-Ting “Alyssa” Peng, yet another former Apple employee who left to join OpenAI, saying, “LOL, I found out I can access the [network storage], so funny,” who replied, “I’m ready.”

Apple also claims that Liu accessed their systems by exploiting an authentication bug using Peng’s work computer after he resigned.

Not exactly the work of criminal masterminds.

The lawsuit also names io, the device startup company founded by, you guessed it, former Apple employees, among them ex-Chief Design Officer Jony Ive, which was acquired by OpenAI in a $6.5 billion deal.

Even without Ive being named in the filing, this feels premeditated. Ive is universally recognised as the co-creator (alongside Steve Jobs) of Apple’s visual identity. The iMac, iPhone, iPad, and Apple Watch are heavily influenced by his design philosophy from the era he helped define.

That probably set off alarm bells at Apple. Loud ones.

So, if OpenAI’s testing the hardware waters with former Apple designers, engineers, and allegedly stolen information, their claim of “building innovative technology” feels more like copying someone else’s homework.

The New York Times ChatGPT Copyright Lawsuit: Different Fight, Same Principle?

While Apple is going after OpenAI from the hardware side, The New York Times is going after them on the data front.

The Times sued OpenAI (and Microsoft) in December 2023, alleging that millions of Times articles were used as training data for AI models, including ChatGPT, without permission. They claim that its copyrighted content was used to build and train commercial AI tools without payment and that those tools can reproduce, summarise, or substitute for human reporting.

Not only is The Times saying its work was copied. It’s also saying that their own work was used to create parasitic products that can directly compete with it, while reducing site traffic, letting people bypass paywalls, and leading to fewer subscriptions and less revenue.

The AI Training Data Argument

Large Language Models (LLMs) are trained on massive datasets so they can “learn” by example.

These datasets include content for websites, online databases, forums, social media posts, and news articles, and AI companies scrape the internet for billions of examples. Some of that content can be copyrighted or under licence, but AI crawlers don’t know that and have a history of ignoring no-crawl directives.

Web scraping isn’t the issue here; search engines have done it for decades to index sites. The issue is what gets scraped, how it is used, and whether permission was given.

AI companies generally argue that the content in training model datasets qualifies as fair use. Fair use is legalese for the limited use of copyrighted material without permission. The line depends on the purpose, transformation (how it is changed from the original to add a new perspective, meaning, etc.), market effect, and the amount.

Publishers see it very differently, saying that AI companies copy unique content to train models that can answer questions without sending readers back to the original source, for example, AI overviews in search engines.

News articles and blogs aren’t just words on a page. They represent hours of writing, research, fact-checking, editing, reviewing, and design. A lot of effort and expertise by a human goes into what you’re reading on your screen.

If an AI model can summarise or reproduce that work without sending traffic back to the publisher, they can kiss their distribution and monetisation goodbye.

The Hidden Evidence

Recently, there’s been a twist in the NYT OpenAI lawsuit. In July 2026, The New York Times and The Daily News claimed OpenAI misled them and the court and accused them of hiding evidence.

The publishers wanted to determine whether their content appeared in OpenAI’s training data and if ChatGPT reproduced it in outputs. The AI giant’s lawyers said they didn’t have the technical capability to search ChatGPT logs and training databases for copyrighted work.

In April 2026, an OpenAI data-privacy engineer squealed under deposition, saying they had already built and run those exact internal searches. Allegedly, this was done using a database of 78 million de-identified ChatGPT conversations to evaluate and track potential copyright infringement.

This directly contradicts what OpenAI has been saying for the last two years.

By that logic, if OpenAI had the means to search its training data and find directly copied outputs, why didn’t they say so? Possibly because they would be shooting themselves in the foot?

Once again, OpenAI denied the allegations, calling them “blatantly false”. It says The Times is trying to access private chat logs and argues that it is defending both user privacy and fair use principles.

The case is now, not just about copyrighted journalism, it is also about if OpenAI can actually prove what was used, what was generated, and what it already knew.

The Link Between the Lawsuits and OpenAI’s Defence

Apple and The New York Times OpenAI class action lawsuit are not making the same claims. But both point to the same thing: OpenAI’s products depend on highly valuable information, and people want to know if it was obtained legally and through the proper channels.

Apple is asking if confidential product information is being used to build new OpenAI hardware, and The Times is asking if copyrighted content helped train models. But the underlying theme is the same, and the courts are going to be asking:

• Where did the data and information come from?
• What was copied, remembered, or reproduced?
• Who should have been paid, credited, or protected?

OpenAI’s defence against Apple will likely focus on the argument that an ex-employee’s skills and experience aren’t the same as stealing trade secrets.

People leaving a company don’t just ‘forget’ everything they learned there when they walk out the door. They can use the knowledge, skills, and experience gained wherever and whenever they choose. The law doesn’t stop people from changing jobs and developing new products in their area of expertise.

The line is crossed when protected information is taken and used without permission. OpenAI will probably say that it stayed on the right side of that line. How they prove it remains to be seen.

Against The New York Times, odds are they’re going to continue flat-out denying they hid evidence, while arguing fair use, privacy, and that they don’t have the technical ability to conduct large-scale internal data analysis.

Now, if only there was an artificially intelligent way to do that…  

Both lawsuits indicate OpenAI’s biggest problem right now isn’t model performance or developing new tech; it’s proving where everything used to build them came from.

What Happens Next

Neither case is simple, and both could shape the next phase of AI regulation and litigation.

The NYT OpenAI lawsuit asks whether content was used without permission to train ChatGPT models. Apple’s lawsuit asks whether OpenAI’s hardware ambitions were built (pun very much intended), at least in part, on confidential information stolen from one of the most secretive companies in technology.

The Apple case could force OpenAI to reveal more about its hardware plans, recruitment practices, and product development process.

The Times case could force OpenAI to disclose more about AI model training data, copyrighted material, chat logs, and testing.

If Apple wins, AI companies could face stricter controls when hiring from competitors, especially in hardware and product design roles.

If The New York Times wins, AI companies could be required by law to license content, document data sources, do more to prevent plagiarised outputs, and compensate publishers and creators.

If OpenAI wins, it could strengthen the company’s position that training data scraping, and new product development can be defended under the guise of fair use and clever hiring.

Either way, some harsh truths will be faced. On a more light-hearted note, the Elon Musk and Sam Altman Twitter (X) war around the Apple lawsuit is absolute comedy gold.

OpenAI built its reputation on making AI models accessible to the public. But the current lawsuits have put the company and its team in the unenviable position of potentially having to explain exactly what goes on behind the scenes.

The OpenAI lawsuits show how important ownership and control have become online. Whether it is protected content, confidential information, or a business name, knowing what belongs to you, matters.

Domains.co.za helps you take control of one of your most important online assets: your domain. With Domain Registration, Web Hosting, professional Email Hosting, and easy website management in one place, you have everything you need to build and protect your online branding.

How to Choose & Register the PERFECT Domain Name

VIDEO: How to Choose & Register the PERFECT Domain Name

FAQS

Other Blogs of Interest

• How AI can help your small business thrive
• Leveraging Generative AI for your small business: Uses and Benefits
• Domains.co.za Introduces South Africa First AI Domain Name Generator
• AI Crawlers Slowing Down Websites: What You Need To Know
• Data Poisoning In AI Models: What It Is And Why You Should Care