Artificial intelligence makers have faced a ton of criticism for borrowing the work of others to train their models. Now, the world’s largest publishers are taking steps to prevent their authors’ work from being plagiarized in the name of progress.
The Bookseller reports that Penguin Random House Publishing has changed the copyright page at the front of its books to address the use of any of its titles as a source for AI training. The text currently reads: “No part of this document may be used or reproduced in any way for the purpose of training artificial intelligence techniques or systems.”
The new language also allows publishers to prevent data absorption by stating that they “explicitly reserve[titles]from text and data mining exceptions.” This part of the revised text is based on the recent directives of the European Parliament on text and data mining exceptions and property rights.
Penguin Random House is the latest publisher to take action against AI model breaches. Earlier this week, the New York Times issued a cease-and-desist letter to AI startup Perplexity, citing its articles and stories that help AI models generate answers for users.
