A large number of Australian books have allegedly been stolen to train artificial intelligence (AI). The Australian Publishers Association (APA) made the claim after a recent search.
About 18,000 titles were found in the now-deleted U.S.-based dataset called ‘Books3’. It pirated (copied without permission) books to train generative AI.
The Australian Society of Authors (ASA), the peak body for writers, says authors “appropriately feel outraged” that their work has been used without their permission.
“This issue is one of basic fairness,” said ASA CEO Olivia Lanchester.
“The inescapable message to authors and artists is that while your work has been essential in developing our product, we’re not prepared to pay you for it.”
Using authors’ titles without permission to develop these technologies means they are “being locked out of the AI boom”, according to Lanchester.
The ASA said it is writing to AI companies to demand action. It’s advocating for the Federal Government to better protect creators’ work.
In the U.S., acclaimed authors such as George R.R. Martin, John Grisham, and Jodi Picoult recently signed on to a lawsuit suing OpenAI, the creators of ChatGPT, for the same reason.
They allege that generative AI models “endanger fiction writers’ ability to make a living, in that [they] allow anyone to generate — automatically and freely (or very cheaply) — texts they would otherwise pay writers to create”.
The lawsuit suggests that companies have instead evaded copyright legislation altogether “to power their lucrative commercial endeavour, taking whatever datasets of relatively recent books they could get their hands on without authorisation”.