Google Just Turned 100,000 Books Into AI Experts — And Publishers Are On Board
For years, the central fight between artificial intelligence and publishing has been about what AI takes from books. This week, a major new experiment flipped the question: what if AI could become part of the experience of owning and reading a book—and publishers actually agreed to it?
On August 27, Google introduced Expert Intelligence, a new initiative that lets readers bring eligible ebooks they purchased through Google Play Books directly into Gemini Notebook. The AI can then answer questions grounded in the book, generate quizzes, infographics and audio overviews, and combine the book with a reader’s own notes or other sources. At launch, Google says more than 100,000 books are supported. Google announcement
A Different Kind of AI-and-Books Story
The timing is striking. Authors and publishers have spent much of the generative-AI era fighting over copyright, training data and whether books were used without permission. Expert Intelligence arrives with a notably different model: Google says it developed the feature with publishers and authors, and readers must own an eligible ebook through Google Play Books before they can use that book as a source inside Gemini Notebook.
If a notebook containing a purchased book is shared, collaborators who do not own that title cannot access its full text or query its contents. Instead, they are directed to purchase their own copy. That ownership gate is one reason the launch is attracting attention inside publishing. The Verge
Why Publishers Are Paying Attention
The launch includes books from Bloomsbury, De Gruyter Brill, Johns Hopkins University Press, Macmillan Publishers, O’Reilly Media and Penguin Random House. Google also worked with more than 15 bestselling authors—including Steven Pinker, Michael Pollan and Jennifer Wallace—on Featured Notebooks that supplement books with additional material. Google
Publishers Weekly described the rollout as an unusual AI application that has drawn positive early reaction from publishing organizations because of its copyright and ownership controls. The publication also noted the contrast with ongoing disputes over how large AI systems have been trained. Publishers Weekly
That distinction matters. The publishing industry is not simply deciding whether it likes or dislikes AI. It is beginning to negotiate what an acceptable commercial relationship with AI might look like.
A Book May Be Becoming More Than a Book
For readers, the most consequential part of the announcement may be how it changes the utility of a purchased ebook. A management book can become an adviser inside a work notebook. A health book can be consulted alongside a reader’s personal notes. A nonfiction title can produce study materials, summaries and audio discussions grounded in the text the reader actually owns.
The Verge reported that users can view the full text of supported books in Gemini Notebook and that Google intends to extend Expert Intelligence beyond books to scholarly articles, magazines and newspapers. Google says the initiative is also expected to reach the Gemini app and AI Mode in Search. The Verge Google
The Business Model Hidden Inside the Reading Tool
There is also a straightforward commercial incentive. Requiring ownership means AI interactions can become a reason to buy a digital book rather than a substitute for buying one. It may also give publishers a new way to keep older nonfiction titles useful long after their initial release cycle.
For authors, that could create something publishing has historically struggled to sustain: an ongoing relationship between a book and its reader after the last page. A title that once sat dormant on a digital shelf can become a reference source a reader repeatedly consults.
What This Could Mean for Authors
The implications extend beyond major publishers. If ownership-gated AI interaction eventually reaches independent and self-published books, authors may begin thinking about books not only as narratives or information products, but as knowledge systems designed to be queried, compared and applied.
That possibility also raises new editorial questions. Writers may need to consider how clearly their arguments can be retrieved by AI, how citations are preserved, what supplemental material belongs in an interactive edition and whether readers will expect a book to function as both a linear work and an intelligent reference source.
The Bigger Shift: AI Moves From Training on Books to Living Beside Them
The most interesting part of Expert Intelligence is not that an AI can summarize a book. That has been possible, legitimately or otherwise, for years. The more important development is the attempt to build a system where the reader owns the source, the publisher participates, the author’s work remains identifiable, and the AI sits on top of that relationship rather than quietly replacing it.
Whether readers embrace the model remains to be seen. But the launch points toward a future in which the fight over AI and literature is no longer only about protecting books from machines. It may also be about deciding what books become when machines are invited in—with permission.
Sources
Image Credit
Header photo: “Boston Public Library Reading Room” by Brian Johnson, via Wikimedia Commons. Licensed under CC BY-SA 3.0. No modifications to the original image. Image source License
Comments