Google is turning reading into a conversation. Its new Expert Intelligence feature lets Gemini Notebook pull any purchased Play Books title, then query, summarize, and remix the content.

Turning Static Text into Interactive Intelligence

Expert Intelligence bridges a personal library and a large language model. In a live demo, Google Labs editorial director Steven Johnson fed Michael Pollan’s Food Rules into the notebook and asked it to draft a week-long recipe plan. The AI produced a complete menu, grocery list, and a new ebook export. He then loaded Kim Scott’s Radical Candor and got a step-by-step script for a tough performance review.

Beyond Q&A, the tool creates infographics, study schedules, and even audio scripts that could become podcasts—all without leaving the notebook.

A Massive Library Supported by Major Publishers

Google secured deals with more than 100,000 titles from publishers such as Penguin Random House, Macmillan, O’Reilly Media, and Johns Hopkins University Press.

The company also rolled out “Featured Notebooks,” where authors like Steven Pinker and Jennifer Wallace curate their own digital spaces. Their notebooks include welcome letters, supplemental readings, and pre-built prompts, giving readers a “digital twin” of the author’s expertise.

Google built strict safeguards to keep the feature a revenue driver, not a replacement. If you share a notebook that references a purchased book, anyone without that book sees only a placeholder that lists the source. The placeholder blocks the text and redirects the viewer to Play Books for purchase.

The design aims to create a “dream team of advisors” while opening new monetization channels for creators.

The Future of Contextual Search and Research

Google plans to extend Expert Intelligence to scholarly articles, magazines, and newspapers. Eventually the same engine will power Google Search’s AI Mode and the standalone Gemini app, letting users pull answers straight from paid journals instead of generic snippets.

Key Takeaways

  • Deep Integration: Gemini Notebook now accesses over 100,000 titles from major publishers via Expert Intelligence.
  • Generative Workflows: Users can spin book content into recipes, management plans, infographics, and AI-generated audio.
  • IP Protection: Shared notebooks reveal text only to recipients who own the source book, prompting others to buy it.

ARTICLE:

Google rolled out “Expert Intelligence” inside Gemini Notebook, letting users pull any Play Books purchase into the AI-driven workspace, ask questions, get summaries and spin out new content while keeping the book’s copyright intact.

The move turns a static ebook into a live data source. In a live demo, Google Labs editorial director Steven Johnson fed Michael Pollan’s Food Rules into Gemini Notebook and asked the system to draft a week-long recipe plan. The AI produced a full menu, complete with grocery lists, that could be exported as a new ebook. A similar run with Kim Scott’s Radical Candor yielded a step-by-step script for handling a tough performance review. The feature goes beyond simple Q&A; it can churn out infographics, study schedules and even audio scripts that could become podcasts, all without leaving the notebook interface.

Why this matters now

Google’s AI models have long been able to browse the open web, but feeding them copyrighted text has been a legal gray area. Publishers have been wary of allowing their catalog to be parsed by large language models (LLMs) because the output could replace sales. By pairing the notebook with Play Books, Google offers a controlled environment: the text is only accessible to the buyer, and any downstream sharing is gated behind a purchase requirement.

The company says the initial rollout covers more than 100,000 titles from major houses such as Penguin Random House, Macmillan, O’Reilly Media and Johns Hopkins University Press. That breadth gives users a sizable knowledge base while giving publishers a new revenue stream tied to AI-enhanced interactions.

How the safeguards work

When a notebook that references a purchased book is shared, recipients who do not own the same title see a placeholder that lists the source but blocks the text. The interface prompts them to buy the book on Play Books before they can query or view it. This design aims to prevent “free-riding” on the original purchase while still allowing collaborative note-taking.

Critics argue that even limited excerpts could be enough for a model to recreate large portions of a work over time. Google counters that the system enforces strict usage limits and that the AI never stores the full text beyond the active session. The company also says it will monitor usage patterns for abuse and can revoke access if it detects systematic scraping.

Google is experimenting with a “digital twin” concept, where authors create curated notebooks around their own books. Notable participants include cognitive scientist Steven Pinker and business writer Jennifer Wallace. Their notebooks contain a welcome note, supplemental readings and pre-built prompts that guide readers toward deeper engagement. For authors, this offers a way to extend the life of a book beyond the printed page; for readers, it provides a guided path to applying the material.

Google hints that the same technology will soon handle scholarly articles, magazines and newspapers. If successful, the ability to ask an AI to pull a specific finding from a peer-reviewed study could reshape academic workflows. The company also plans to weave the Expert Intelligence engine into the AI-focused mode of its search product and the standalone Gemini app, meaning that the next time a user asks a web search for “the latest data on renewable energy storage,” the answer could be drawn directly from a paid journal article rather than a generic snippet.

Potential downsides

  • Cost to creators: Even with safeguards, some authors worry that AI-generated derivatives could dilute the market for their original work.
  • User privacy: The notebook logs queries and generated content, raising questions about how that data is stored and used.
  • Access inequality: Users without the means to purchase every referenced title may hit roadblocks, limiting the collaborative potential of the platform.

What to watch

  • Publisher response: If major houses see a sustained revenue lift, they may broaden the catalog; if not, they could pull back.
  • Regulatory scrutiny: Copyright agencies worldwide are still defining how AI can legally interact with protected text.
  • Feature rollout: Google has not disclosed a timeline for extending the service beyond books, but updates are expected later this year.

Key takeaways

  • Deep integration: Gemini Notebook now supports over 100,000 Play Books titles via Expert Intelligence.
  • Generative workflows: Users can turn book content into recipes, management plans, visual aids and AI-generated audio.
  • IP protection: Shared notebooks only reveal text to recipients who also own the source book, directing others to purchase.

The experiment puts a premium on balancing AI convenience with creators’ rights; its success will likely set the tone for how generative models handle copyrighted material across the web.