Penguin Random House amends copyright rules to say no to AI training: What does this mean

AI (artificial intelligence) companies have faced a lot of criticism from artists, developers, companies and authors from across the globe for training their AI models on the copyrighted data. Now, one of the biggest publishing companies in the world, Penguin Random House (PRH), has taken steps to protect its authors and ensure that their works isn't used by AI companies for training their AI models.

According to a report by The Bookseller (via Engadget) Penguin Random House has amended the copyright rules printed at the front of its books to prevent AI companies from using the content from its books for training various AI models. "No part of this book may be used or reproduced in any manner for the purpose of training artificial intelligence technologies or systems," the updated wording now says.

In addition to this, the company has also added special provisions in its copyright rules that prevent AI companies from mining the text in its book for their use. "In accordance with Article 4(3) of the Digital Single Market Directive 2019/790, Penguin Random House expressly reserves this work from the text and data mining exception," the company says in the amended section of the front page of its books.

Furthermore, the report says that these changes will be included in all new titles and reprints that the company publishes across the globe and they are in line with the European Parliament's directive on text and data mining exceptions and ownership.

For the unversed, EU parliament's directives include "a mandatory exception allowing research organizations to make reproductions and extractions in order to carry out TDM of works or other subject-matter to which they have lawful access for the purposes of scientific research". To put into perspective, research organisations can use work of authors for text and data mining for research purposes only if they have relevant permissions fro the publisher.

What does this mean?

With this, Penguin Random House has become the first major publisher to modify its copyright rules to reflect on the rapid development of AI models and their adoption. It also marks a firm stance by the company on AI models being trained on text from books without relevant permissions.

It is worth noting that not all publishers have taken a stand against AI companies using the work of their authors for training their AI models. Some publishers such as Wiley, Taylor & Francis and Oxford University Press have partnered with AI companies giving them right to use their authors' works for training their large language models or LLMs.