Australian authors are furious over a recent Productivity Commission (PC) interim report that says AI could deliver a $116 billion boost to Australia’s economy over the next 10 years.<\/p>\n\n\n\n
The centrepiece of the report was a proposal to implement a text and data mining (TDM) exception to the Copyright Act, which would permit tech companies to use copyrighted work to train AI.<\/p>\n\n\n\n
In July, former Atlassian CEO Scott Farquhar made a similar suggestion<\/a> in his address to the National Press Club, arguing that a TDM exception could “unlock billions of dollars of investment in Australia”.<\/p>\n\n\n\n Your ultimate guide for staying in the know and getting inspired. An essential wrap-up of arts news and the best interviews, brought to you by the Arts team at ABC Radio National. <\/p>\n\n\n\n Lucy Hayward, CEO of the Australian Society of Authors (ASA), said the proposal gave “a free pass” to multinational tech companies, such as Google, Meta and OpenAI, to continue using unauthorised copyrighted material to train their AI models.<\/p>\n\n\n\n “Why should we create a situation where billion-dollar tech companies can profit off authors’ work, but not the creators who made the work? It’s an entirely absurd proposition,” Hayward told ABC Arts.<\/p>\n\n\n\n While the government has yet to deliver a formal response to the interim report, Arts Minister Tony Burke stated that the unauthorised use of copyrighted material for commercial purposes constituted theft.<\/p>\n\n\n\n “We have copyright laws,” the minister said in a speech at the 2025 BookUp conference In Sydney.<\/p>\n\n\n\n “We have no plans, no intention, no appetite to be weakening those copyright laws based on this draft report.”<\/p>\n<\/blockquote>\n\n\n\n Tech companies have already used<\/a> unauthorised copyrighted material to train AI platforms.<\/p>\n\n\n\n In March, The Atlantic published a tool<\/a> that made it possible to search the LibGen database, an online trove of pirated books and academic papers that Meta used to train its generative AI language model.Authors furious Meta ‘plundered’ their work to train AI<\/a><\/p>\n\n\n\n Tech giant Meta used pirated books by Australian authors including Hannah Kent and Charlotte Wood to train its AI language model.<\/p>\n\n\n\n It followed similar revelations in 2023 that a database of pirated material known as Books3 had been used to train Meta’s AI model Llama, Bloomberg’s BloombergGPT and EleutherAI’s GPT-J.<\/p>\n\n\n\n The work of countless Australian authors appeared in the pirated databases, including Charlotte Wood, Tim Winton, Helen Garner and Richard Flanagan.<\/p>\n\n\n\n A TDM exception would allow this type of use of copyrighted material without compensating the author or seeking their consent.<\/p>\n\n\n\n\n
Sanctioning an illegal act<\/h2>\n\n\n\n