AI Firms Book Destruction Practices Fuel Copyright Debate

AI companies are facing criticism for purchasing, scanning, and destroying books to train language models, reigniting debates over copyright, fair use, and authors' rights.
on Jul 15, 2026

Copyright Issues Arise Due to Destruction of Books by AI Companies

With the fast development of artificial intelligence technology, a debate has arisen after some reports claimed that some artificial intelligence companies had purchased and destroyed physical copies of books in order to make digital copies for training purposes for their language model.

Books That Have Been Acquired, Digitized, and Then Destroyed

Based on recent reports, some of the AI companies have bought thousands of paper-based books, stripped them of their covers, and then digitized all of the pages before destroying the books themselves. These companies justify this practice by claiming that since the books have been legally acquired, they can be used to create data sets to develop AI.

However, authors and publishers claim that although the books have been legally acquired, turning them into data sets to create AI is far from what was traditionally allowed.

Publishers and Authors Raise Objections

The revelations have prompted renewed criticism from authors, publishers, and copyright advocates, many of whom argue that their creative works are being used to build profitable AI models without permission or compensation. Industry groups believe such practices undermine intellectual property rights and could set a concerning precedent for the future of publishing.

Several lawsuits involving AI developers and copyright holders are already underway, with courts expected to clarify how existing copyright laws apply to AI training.

Growing Legal and Ethical Dilemma

Advocates for developing AI contend that training algorithms with massive amounts of text data is crucial for language technology development, and purchasing physical books serves as a legitimate means of obtaining text data. However, critics believe that digitization and the use of books for machine learning are different from traditional reading and should be subject to licensing agreements.

With increasing popularity of AI in various industries, it becomes evident that there is an urgent need to reform the existing copyright system to address technological innovations while providing sufficient protection to authors and publishers.

The result of legal disputes might have implications for how AI firms can acquire copyrighted materials in the future, which makes this dilemma one of the most discussed topics in the book publishing industry.