Apple faces new AI lawsuits

Apple’s AI legal problem started with a familiar complaint. The company showed off a much smarter Siri at WWDC in June 2024, tied that vision to the iPhone 16, then failed to ship the headline features on time. That gap produced consumer lawsuits in March and April 2025, including *Landsheft v. Apple* in federal court in Northern California, which accuses Apple of false advertising and unfair competition over Siri features that were promoted but not available when buyers paid for new phones (courtlistener.com, macrumors.com). Then the problem widened. In June 2025, shareholders sued in *Tucker v. Apple*, arguing that Apple also misled investors about how ready its AI work really was. That case says the company downplayed how long it would take to integrate advanced AI into Siri, and that the truth only became obvious after Apple delayed the upgrade and then spent WWDC 2025 talking around it. Reuters reported that the complaint covers a period ending on June 9, 2025, when Apple’s keynote offered only modest AI updates and disappointed analysts who had expected more (courtlistener.com, tech.yahoo.com). That was already bad enough. What Macworld pointed out this week is that Apple is now getting squeezed from the other side too. The newer cases are not about shipping delays at all. They are about what Apple may have done to build its AI in the first place, and what rules it had in place while doing it (macworld.com). The clearest example is the copyright fight over training data. In October 2025, neuroscientists Susana Martinez-Conde and Stephen Macknik sued Apple in Northern California, alleging that Apple used pirated books from the Books3 dataset and other “shadow libraries” to train Apple Intelligence models, including OpenELM. The case was filed as *Martinez-Conde v. Apple Inc.* and put Apple in the same trench as Meta, OpenAI, and other AI companies already defending how they built their models (courtlistener.com, fastcompany.com, mashable.com). That matters because it changes the shape of Apple’s exposure. The Siri-delay suits say Apple promised too much. The copyright suit says Apple may have taken too much. One set of claims is about marketing. The other is about governance. Together they suggest that Apple’s AI risk is no longer just a product-timing mess. It is becoming a question of whether the company had real controls over what its models were trained on, what it told customers, and what it told Wall Street (macworld.com, apple.com). That shift is especially awkward for Apple because the company has spent years selling itself as the careful one. Its AI pitch leaned on privacy, restraint, and on-device processing. But courts do not grade branding. They look at records, timelines, and who knew what when. In the shareholder case, named defendants include Tim Cook, current CFO Kevan Parekh, and former CFO Luca Maestri. In the consumer case, plaintiffs say they paid a premium for an iPhone 16 partly because Apple’s ads made the missing Siri features sound real and near-term. In the copyright case, the dispute turns on whether Apple’s model training drew from a pirated library containing more than 190,000 books (tech.yahoo.com, macrumors.com, mashable.com).