Apple is facing a growing legal challenge over its use of copyrighted materials to train its artificial intelligence models. The tech giant has been hit with a copyright infringement lawsuit alleging that it used books in AI model training without permission. This lawsuit, filed in the US District Court for the Northern District of California on October 9, 2025, by neuroscientists Susana Martinez-Conde and Stephen Macknik, marks the second such legal action against Apple in just over a month.
The plaintiffs, professors at SUNY Downstate Health Sciences University, claim Apple illegally accessed thousands of pirated books from "shadow libraries" to train its Apple Intelligence AI. Martinez-Conde and Macknik specifically allege that Apple used their copyrighted works, including "Champions of Illusion: The Science Behind Mind-Boggling Images and Mystifying Brain Puzzles" and "Sleights of Mind: What the Neuroscience of Magic Reveals About Our Everydayals," in the AI training process. They claim that Apple trained its OpenELM model using a pirated database called Books3, a shadow library containing over 190,000 works.
This lawsuit follows a similar complaint filed in September by authors Grady Hendrix and Jennifer Roberson, who accused Apple of using its web-crawling software, Applebot, to scrape copyrighted material from "shadow libraries". These online collections contain unlicensed copies of copyrighted books.
The plaintiffs in the recent lawsuit argue that Apple's market value surged by over $200 billion following the unveiling of "Apple Intelligence," suggesting the company directly benefited from the alleged copyright infringement. They are seeking unspecified damages and a court order to prevent Apple from using copyright-infringing materials in its AI training.
Apple is among several tech companies facing legal battles over the use of copyrighted material in AI training. Meta, Microsoft, and OpenAI have also been targeted with similar lawsuits. These companies often invoke the "fair use" legal doctrine, arguing that using copyrighted material for AI training without permission or payment is permissible. However, the courts have not always agreed, and the issue remains a contentious one.
In a similar case, Meta was sued for using data from shadow libraries to train its Llama AI, but a US district judge ruled that the use was considered "fair use". In contrast, Anthropic recently agreed to a $1.5 billion settlement with authors who claimed its Claude chatbot was trained on their works without authorization.
These lawsuits highlight the growing tension between copyright law and the data-intensive nature of AI development. Experts suggest that tech companies may need to prioritize compliance by entering into licensing agreements with authors, publishers, and content owners to legally obtain materials for model training. However, this could significantly increase the cost of AI development and potentially impact the pace and accessibility of future AI innovation. Some reports indicate that Apple's technical reports lack transparency regarding how it filters pirated copies or avoids shadow libraries. This lack of clarity could increase compliance and liability risks for the company.
