AI Wants Our Books. Australian Authors Are Asking Who Gave Permission

A group of prominent Australian writers has put a deceptively simple question to lawmakers: if an AI company wants to use our books, why shouldn't it ask first? At a parliamentary hearing in Sydney on 21 October, authors Anna Funder, Charlotte Wood and Markus Zusak joined publishing representatives to argue against weakening copyright protections for artificial intelligence. The story was reported on 22 October by ArtsHub. For authors watching the technology develop from the UK, it is a reminder that the argument about AI is no longer confined to what a chatbot can write. It is about who controls the work that helped make it possible.

A hearing with real consequences

The Australian Joint Select Committee on Artificial Intelligence heard from writers and publishing organisations who want consent, transparency and fair payment at the centre of any response to AI and copyright. Alongside the novelists were Australian Society of Authors chief executive Lucy Hayward, Australian Publishers Association representatives and Hachette Australia chief executive Louise Stark. Their presence matters: this is not simply a handful of individual writers protesting new technology. It is an industry asking lawmakers to consider how the economics of original writing survive in an AI-driven market.

We should be precise about what happened. A parliamentary hearing is an opportunity to give evidence and influence policy, not the passage of a new law. Nor does opposition to weakened copyright rules necessarily mean opposition to every use of AI. The central issue is whether rights holders should have a meaningful say when their work is used for commercial training, and whether they can discover and negotiate that use.

The difference between access and permission

There is an awkward assumption lurking beneath much of the AI debate: if a book, article or illustration can be accessed digitally, perhaps it can also be treated as raw material for training. But being able to read a work is not the same thing as owning the rights to exploit it commercially. Authors make their living, when they can, by controlling particular uses of their work. Copyright is imperfect, and licensing can be complicated, but permission is not a quaint obstacle that technology has made irrelevant.

Developers have arguments of their own. They contend that training systems learn patterns rather than simply distributing copies, and that broader access to data can support useful innovation. Those claims deserve serious examination rather than slogans. Yet they do not, by themselves, settle questions of lawful access, compensation or whether a system's outputs could substitute for the material on which it was trained. Different jurisdictions answer those questions differently, and courts are still working through them.

Why this matters to UK authors

Although the hearing took place in Australia, British writers face much the same practical uncertainty. The UK debate over text and data mining, licensing and opt-out systems has made one thing clear: technical convenience and creative consent are not interchangeable. In June, Publishers' Licensing Services said more than 250 UK publishers had opted into a collective AI licensing scheme. That is evidence of an emerging commercial route, but it is not a universal answer for every author or every contract.

An author should be able to ask who has authority to license a book, what uses a licence covers, whether the author receives a share and what happens to older works whose rights arrangements were drafted long before generative AI existed. A publisher's ability to negotiate a deal does not automatically tell us whether an individual writer's interests have been protected. Those details are where the principle of fair payment either becomes real or evaporates.

AI assistance is not the same as surrendering a manuscript

As a science-fiction author, I find AI fascinating. I can use it to challenge an outline, organise research, explore marketing ideas or help visualise a fictional world without asking it to take over the creative decisions that make the story mine. I want authors to have room to experiment responsibly. I also want them to retain control over whether their published books or unpublished manuscripts become training material for somebody else's product.

Those are different choices. A writer deciding to use an AI tool for administrative support is not giving blanket consent for a third party to ingest a novel. Equally, a writer who asks a model to generate substantial narrative prose has a different disclosure and editorial responsibility from someone using it to brainstorm chapter headings. Treating every interaction with AI as identical makes it harder to build rules that are fair, practical and intelligible.

What I would do now

First, read your publishing and service contracts, especially clauses covering subsidiary rights, electronic exploitation, data mining and future technologies. If you are unsure what rights you have granted, ask your agent, publisher or a qualified adviser. Second, establish what happens to a manuscript before uploading it to any AI service: whether the text is retained, whether it can be used for training and whether the provider offers controls that meet your needs.

Third, keep your own working drafts and editorial notes. Not because authors should have to prove their innocence to an unreliable detector, but because good records are useful in an ordinary professional writing practice. Finally, decide where your personal boundary sits between human-led creation, AI-assisted administration and material machine-generated contributions. You will make better decisions if you can explain that boundary before a dispute arises.

A bigger question than one country's copyright rules

The Australian hearing is one chapter in a wider argument about who benefits from the work that feeds AI systems. The answer should not be reduced to a choice between protecting authors and permitting innovation. We need both. The difficult work is designing permissions, licensing arrangements and transparency obligations that ordinary writers can actually understand and use.

For me, the principle remains straightforward: a creative industry cannot thrive if its creators are expected to provide the raw material while having no meaningful voice in the terms. New technology deserves the chance to prove its value. Authors deserve the chance to decide how their work participates in that future. What would a genuinely fair AI licensing agreement look like to you?

Sources and further reading

ArtsHub, 22 October 2026 — Australian authors fight the watering down of copyright law for AI: https://www.artshub.com.au/news/news/australian-authors-fight-the-watering-down-of-copyright-law-for-ai-2877742/

The Bookseller, 24 June 2026 — Over 250 UK publishers opt in to collective AI licensing scheme: https://www.thebookseller.com/news/over-250-uk-publishers-opt-in-to-collective-ai-licensing-scheme

Pinsent Masons, March 2026 — Drop plans for AI-related copyright exception, UK ministers urged: https://www.pinsentmasons.com/out-law/news/drop-plans-ai-related-copyright-exception-uk-ministers

About Rob Frankson: https://www.121minutes.com/about

Subscribe to AI & The Author: https://www.121minutes.com/subscribe

AI & Editorial Transparency

AI & The Author is edited and published by Rob Frankson. Artificial intelligence is used to assist with news research, initial drafting, content organisation and supporting imagery. All articles are reviewed and, where necessary, edited by Rob Frankson before publication. The opinions, editorial position and final decision to publish remain the author's.