The use of generative artificial intelligence (AI) in writing books is changing how people view and value literature. In the world of publishing, the role of AI is becoming more important, especially since the European Union passed the AI Act in August 2023. This law has led to new labeling practices, where books may now include information about whether AI was used in their creation. These labels affect both the material aspects of a book—like its design—and its symbolic value, which relates to how readers perceive its authenticity and authorship. In France, AI has become a common but often unspoken tool in many businesses. By 2025, 18% of companies with at least 10 employees used AI, up from 6% in 2023. Publishing houses are no exception. According to recent surveys, book professionals often use AI for tasks like writing emails, summarizing documents, or creating summaries. Tools like Insight, which is marketed as an AI that assists editors, are helping to improve metadata and marketing materials. These tools are also being discussed in reports by the National Assembly, which highlights the growing use of AI across cultural sectors. Despite this widespread use, the public is rarely informed about AI's role in the production of books. However, when AI is involved in the actual creative process—like writing the story—readers expect transparency. Authors who use AI are often asked to mention it, as readers want to know how the book was made. AI is generally accepted when it's used for non-creative tasks, but it becomes a point of discussion when it's involved in the actual writing of a book. To address these concerns, some publishers have started labeling books to indicate whether AI was used. In France, labels like "Human Creation" and "Guaranteed without artificial intelligence" are being used to show that a book was written without AI. In the U.S., the Authors Guild has introduced a "Human Authored" certification. Other publishers choose to be more open by clearly stating the use of AI on the cover or in the back matter. For example, the novel If Rome Had Not Fallen includes a label stating it was written and illustrated with AI, which helps promote the book as a novelty. In contrast, the Japanese novel Tokyo Sympathy Tower does not mention AI in its French edition, despite the author's admission of its use. These varying approaches show that transparency is not just about whether information is included, but also how it is presented. Some publishers highlight AI use as a feature, while others avoid mentioning it. This diversity of strategies reflects the broader trend of AI normalization, which includes new European regulations requiring AI-generated content to be labeled to prevent deception and build user trust. While these rules do not specifically apply to books, the publishing industry is responding by experimenting with different ways to inform readers. This shift also raises deeper questions about the nature of books themselves. As Gérard Genette, a literary theorist, once wrote, the way a book is presented—its cover, title, and preface—shapes how readers approach the text. If AI is now part of this "paratext," it could change how books are interpreted and valued. The challenge for publishers is not just to inform readers about AI use, but to decide where and how this information should be placed, and what role it plays in the reader’s experience. This evolving landscape invites further research into how these changes affect the way books are created, read, and valued.