Apple Sued Over AI Training: Did the Company Steal Books

Case Overview: A class action lawsuit alleges Apple copied copyrighted books without permission to train its Apple Intelligence platform, including its Foundation Intelligence and OpenELM language models.

Consumers Affected: U.S. copyright holders whose works were used in Apple Intelligence training.

Court: U.S. District Court for the Northern District of California, San Francisco Division

Apple copied copyrighted books without permission to train its Apple Intelligence platform, including its Foundation Intelligence and OpenELM language models, a new class action lawsuit alleges.

The lawsuit says Apple secretly relied on pirated datasets like Books3, which contain nearly 200,000 books, to give its models high-quality training material without licensing the works or paying the authors. By doing so, Apple allegedly created AI outputs that compete directly with the very books it copied, eroding their value and undercutting sales.

Plaintiffs Say Their Works Were Taken Without Permission

The case was filed by authors Grady Hendrix of New York and Jennifer Roberson of Arizona. Both say their published novels, My Best Friend’s Exorcism and Sword-Bound, were pulled into the Books3 dataset and used to train Apple’s AI models without their knowledge.

The complaint argues that Apple not only copied these works but also built a permanent internal library of pirated books to fuel future AI training. That, the authors say, leaves them cut out of a rapidly growing licensing market for training data, while Apple reaps the commercial rewards.

Lawsuit Alleges Apple’s AI Outputs Compete with Original Books

Apple Intelligence was unveiled in 2024 as a breakthrough technology to enhance its devices, from iPhones to Macs. The company described it as a “new era” of personal AI, touting its ability to write, summarize, and reason. The announcement alone added more than $200 billion to Apple’s market value in a single day.

But plaintiffs argue the foundation of that success rests on stolen work. By copying books wholesale, Apple allegedly denied authors both payment and control. The lawsuit warns that AI-generated outputs could flood the marketplace with cheap knockoffs and “companion” texts that siphon sales from genuine titles, diluting authors’ recognition and royalties.

According to the filing, Apple trained its OpenELM models on RedPajama, a dataset that directly incorporates Books3. That collection was taken from a shadow library called Bibliotik and contains nearly 200,000 books stripped into text files. Apple’s Foundation Language Models, the core of Apple Intelligence, were trained on the same data, the suit says.

The complaint alleges Apple knew it was using copyrighted works and concealed the sources of its training sets to dodge legal scrutiny. While Apple struck a licensing deal with Shutterstock for images, it allegedly made no effort to license books from authors or publishers.

Apple Joins Tech Firms Facing Legal Battles Over AI Data Use

Apple isn’t alone in facing scrutiny. Transcription app Otter.ai is fighting claims that it secretly recorded and analyzed conversations to improve its machine learning. 

Amazon is in court over Alexa devices allegedly capturing voice snippets and location data without consent. Meanwhile Dialpad, a California tech firm, is accused of recording consumer calls for AI training without disclosing it to users. These cases reflect a broader reckoning over how AI companies collect and use data, and whether individuals have any say in the process.

The proposed class would cover all U.S. copyright holders whose works were used in Apple Intelligence training. The plaintiffs are suing for direct copyright infringement, seeking damages, interest, fees, and other costs. Put simply: the authors say Apple built its multi-billion-dollar AI on the backs of their books and now they want Apple to pay up.

Case Details

• Lawsuit: Hendrix, et al. v. Apple Inc.
• Case Number: 3:25-cv-07558
• Court: U.S. District Court for the Northern District of California, San Francisco Division

Plaintiffs' Attorneys

• Benjamin Gould, Derek W. Loeser, Chris N. Ryder, William K. Dreher, and Elizabeth W. Tarbell (Keller Rohrback L.L.P.)
• Matthew Butterick (Butterick Law PC)

Do you have concerns about how anything you’ve published online is used by companies with AI products? Let us know in the comments section below!