In a major new development in the battle over intellectual property in the artificial intelligence age, Apple has been sued by authors who allege that the company illegally used their copyrighted books to train its AI systems. The proposed class-action lawsuit, filed Friday in the U.S. District Court for the Northern District of California, accuses Apple of copying protected works without consent, credit, or compensation.
The lawsuit represents another significant step in a growing legal and ethical fight over how large language models (LLMs) the backbone of today’s generative AI tools acquire their vast knowledge.
The lawsuit, brought by authors Grady Hendrix and Jennifer Roberson, centers on Apple’s “OpenELM” large language models. According to the complaint, Apple allegedly used a “known body of pirated books” to train these models, effectively building a potentially lucrative AI platform on the unpaid labor of writers.
“Apple has not attempted to pay these authors for their contributions to this potentially lucrative venture,” the lawsuit states, alleging that Hendrix’s and Roberson’s works were among those illegally copied and included in the training datasets.
The lawsuit seeks damages and other relief on behalf of a proposed class of authors whose works were allegedly misappropriated.
A Pattern Across the AI Industry
Apple is not the first major technology company to face this type of litigation. The case adds the iPhone maker to a growing list of corporations now embroiled in copyright-related disputes involving AI.
- Anthropic: Just hours before the Apple lawsuit became public, Anthropic disclosed that it had agreed to pay $1.5 billion to settle a class-action lawsuit brought by authors accusing the company of using pirated books to train its chatbot, Claude. Though Anthropic did not admit liability, plaintiffs’ attorneys called the settlement the largest copyright recovery in history.
- Microsoft: In June, Microsoft was sued by a group of authors claiming that their books were used without permission to train its Megatron AI model.
- OpenAI and Meta: Both companies have also been hit with lawsuits over allegedly using copyrighted materials in training datasets without licenses or compensation.
The Apple lawsuit appears to follow a similar legal theory: that massive datasets, often sourced from the internet, contain copyrighted works used without consent, violating U.S. copyright law.
The Legal Questions: Fair Use or Infringement?
At the center of all these cases lies one crucial legal question: does using copyrighted works to train AI models constitute fair use?
Fair use is a legal doctrine that allows for the use of copyrighted materials without permission in certain contexts, such as commentary, criticism, research, or education. Technology companies have argued that training AI models transforms the original works into something new, a statistical model capable of generating text, not reproducing books word-for-word.
However, plaintiffs argue that simply copying their works into datasets especially from pirated sources goes beyond what fair use allows. Courts have issued mixed rulings so far:
- In one high-profile case, Judge William Alsup ruled that Anthropic’s use of books to train Claude could qualify as fair use, but found the company violated authors’ rights by storing millions of pirated works unnecessarily in a “central library.”
- Another federal judge, presiding over a lawsuit against Meta, suggested that using copyrighted works without permission for AI training could be unlawful in “many circumstances.”
The lack of clear, consistent rulings means that cases like Apple’s could shape the future of AI development and copyright law.
The Stakes for Apple
For Apple, the lawsuit could have far-reaching implications beyond potential financial damages. Apple is still catching up in the AI race, and its AI efforts including the OpenELM models and a planned overhaul of its digital assistant, Siri — are considered critical to its future competitiveness.
If courts impose strict liability on AI companies for the use of copyrighted works, it could:
- Increase development costs significantly, as companies would need to license training data.
- Slow innovation by forcing AI firms to rely on smaller, vetted datasets.
- Lead to further lawsuits and settlements, draining resources and creating uncertainty around product roadmaps.
Given the sheer scale of AI training datasets, which often involve billions of tokens scraped from the internet, retroactively identifying and compensating rights holders could be extraordinarily complex.
A Turning Point for Creators and AI Developers
For authors and other creators, lawsuits like this represent a chance to assert control — and demand compensation — over how their intellectual property is used in the age of machine learning.
Mary Rasenberger, CEO of the Authors Guild, recently called the Anthropic settlement “a vital step in acknowledging that AI companies cannot simply steal authors’ creative work to build their AI.” While she did not comment on the Apple lawsuit specifically, the sentiment underscores a growing consensus among creators: AI companies should not be exempt from the rules that protect intellectual property.
At the same time, some legal scholars warn that overly restrictive rulings could hinder AI innovation, limiting the ability of U.S. companies to compete globally in a field dominated by fast-moving advances in China, Europe, and elsewhere.
The lawsuit against Apple is in its early stages. The plaintiffs are seeking class-action status, which could expand the scope of the case to include thousands of authors if certified. Apple has yet to comment publicly on the matter but is expected to mount a vigorous defense, likely relying on fair-use arguments and technical distinctions about how its OpenELM models operate.
Meanwhile, settlements like Anthropic’s suggest that some companies may prefer to pay rather than risk a trial with potentially massive damages and uncertain legal precedent. Whether Apple follows that path or fights to the end could influence how the entire tech industry approaches AI training in the future.
The Apple lawsuit underscores a larger truth: the rules governing AI and intellectual property are still being written, often in real time, by judges facing unprecedented questions. As generative AI continues to reshape industries, courts, lawmakers, and technology leaders will need to balance innovation with the rights of creators.
For now, authors like Hendrix and Roberson are asking a simple question with profound implications: who gets to profit from their words in the age of artificial intelligence? The answer could redefine the economics of both publishing and AI for years to come.
