AI That Makes You Buy the Book First: A Better Deal for Authors?
- Rob Frankson
- 10 hours ago
- 4 min read
For years, the argument about AI and books has focused on what technology companies took from authors: copyrighted text, scraped libraries and the unresolved question of whether writers should have been asked or paid. A new development this week points towards a different model — one in which the book remains a product the reader actually has to own.
Quick Shortcuts
A small change with a very big idea behind it
On 27 August, Google announced Expert Intelligence for Gemini Notebook. The feature lets readers add selected ebooks they have purchased through Google Play Books to an AI notebook and then question, analyse and work with the book as a trusted source. Google says the launch covers more than 100,000 books from publishers including Bloomsbury, Macmillan, O’Reilly Media, Penguin Random House and Johns Hopkins University Press.
The interesting part for authors is not simply that an AI can answer questions about a book. We have been able to upload documents to AI systems for some time. What matters is the commercial rule wrapped around the feature: the reader must own an eligible copy. If the notebook is shared with someone who does not own that book, the other person is prompted to buy their own copy before they can interact with it.
That may sound like a technical detail, but it is actually a potentially important principle: AI access does not automatically have to mean free access.
Why writers should care
Much of the anxiety surrounding generative AI and publishing comes from the feeling that the normal relationship between creator, reader and payment has been broken. Books were ingested into models at enormous scale, often without authors knowing exactly what had happened to their work. The resulting systems could then generate text, summaries or stylistic imitations without the original writer necessarily receiving anything.
Expert Intelligence suggests another route. Instead of treating the book as raw material that disappears invisibly into a model, the book remains an identifiable source. The reader owns it. The AI works with it. The answers can be grounded in it. And access can still be tied to a legitimate purchase.
That does not solve the wider copyright argument, and it certainly does not answer every question about AI training. But it demonstrates something important: technology companies can design AI products around rights and ownership when they choose to.
Could AI actually make a purchased book more valuable?
There is another side to this that writers should not ignore. AI may eventually increase the usefulness of a book after it has been bought. A reader could question a complex non-fiction title, generate a study guide, compare its ideas with their own notes or create an audio overview. Google is explicitly positioning the feature as a way to keep engaging with an author’s expertise rather than reading a book once and putting it back on the shelf.
For publishers, that creates an intriguing possibility. The e-book stops being merely a digital copy of printed pages and becomes a licensed body of knowledge that a reader can actively explore. If implemented fairly, that could make legitimate ownership more attractive rather than less.
There are risks. Authors and publishers will want to know exactly how their work is used, whether AI interactions substitute for reading, how royalties and licensing develop, and whether the system ultimately helps readers discover more books or simply keeps them inside one technology company’s ecosystem. Those questions deserve scrutiny.
AI-assisted is still not AI-authored
For me, this reinforces a distinction that is becoming increasingly important. Using AI around a creative work is not automatically the same thing as asking AI to create that work. I write my fiction. The characters, worlds, narrative decisions and finished stories are mine. AI can be useful around that work , for visualisation, marketing, research support, online content and some of the practical business of being an independent author. Without becoming the author of the books themselves.
The same distinction can apply at the other end of the publishing chain. A reader using AI to interrogate a book does not necessarily diminish the author. If the system respects ownership, identifies the source and encourages legitimate purchase, AI could become another interface between the writer’s ideas and the reader.
The bigger shift: from extraction to permission
This week publishers are again discussing AI, copyright and trusted content, while writers continue to worry about synthetic books, imitation and undisclosed machine-generated manuscripts. Against that backdrop, a system built around “you need to own the book first” feels surprisingly significant.
The future of AI and publishing does not have to be a choice between banning the technology and surrendering everything to it. There is a third possibility: build useful AI services around clear provenance, permission, ownership and human authorship. That is a much healthier conversation for writers than simply asking whether a machine can produce another novel.
The question for authors
If AI can make a book more useful while still requiring the reader to buy and own it, could this become a model authors should support or does allowing AI inside the book create a new set of risks we have not yet understood?
Sources
About Rob Frankson
Rob Frankson is a science-fiction author and creator of the Near Galaxy Saga. His work explores humanity, technology and the consequences of the choices we make. AI & The Author follows the changing relationship between artificial intelligence, writers, publishing and the creative business of being an author.
Join the conversation
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.
.png)


Comments