The March 2026 London Book Fair had one storyline that swallowed the rest: artificial intelligence and copyright. Per Publishing Perspectives, sessions grappled with the Stability AI case and GEMA v. OpenAI, floods of AI-generated titles, and a trade trying to work out what its output is worth when machines can imitate it.
Per Publishing Perspectives' coverage of the March 2026 fair, panels tracked the twin court cases now anchoring the debate — the Stability AI litigation and GEMA v. OpenAI over training data — because their logic travels straight to books. One session title captured the mood plainly: "Mission Impossible? Protecting Authors' Rights in the Age of AI," featuring leaders from authors' organisations. The question is no longer whether AI changes publishing, but who pays for the change.
AI copyright will be a defining story for publishing through the rest of 2026, as Ulverscroft put it, per Publishing Perspectives.
The anxiety has numbers behind it. Per the National Centre for Writing, 51% of published novelists agree that AI is likely to displace their work. And at the fair itself, the Society of Authors launched its "Human Authored" labeling scheme, per a March 2026 Society of Authors announcement — a stamp aimed at readers who want to know that a person, not a model, wrote the book in their hands.
What does the AI focus change for readers?
Not much this year, potentially a lot later. The fights over training data and licensing will shape which books get published, how much they cost, and whether "written by a human" becomes a label you can check. Watch for human-authored badges and cleaner search results as the near-term outcomes.
In the United States, the U.S. Copyright Office's AI initiative is running its own review of how copyright applies to AI inputs and outputs, and whatever it concludes will echo through English-language publishing. Meanwhile, licensing talks between publishers and AI firms have become routine even as lawsuits grind on, meaning the price of training data is being set twice — once in court, once in contracts. Readers will feel it in catalog decisions before they feel it anywhere else.
Why is discoverability the overlooked story from the fair?
Because the lawsuits are loud and the search bar is quiet. "AI slop" — machine-generated books mass-produced for clicks and keywords — surfaced repeatedly in fair programming, per Publishing Perspectives, and it degrades the reader's most basic tool: search. When storefronts fill with lookalike titles, finding a good book gets harder, ratings get noisier, and the cost lands on readers and honest authors alike. Courts will take years; the slop is a today problem.
Which AI court cases were discussed at the fair?
Per Publishing Perspectives, sessions tracked the Stability AI litigation and GEMA v. OpenAI, two cases testing whether training on copyrighted works without permission is lawful. Both matter to publishing because their logic extends to books. The outcomes will shape licensing norms and, ultimately, what AI companies must pay for.
What does AI slop mean for readers?
Machine-generated books mass-produced to chase keywords and trends. The problem is not one bad book but thousands of them, which crowd search results, inflate review counts, and make genuine recommendations harder to trust. It is a discoverability problem before it is a quality one — and readers pay the search tax.
What did the survey of novelists actually find?
Per the National Centre for Writing, 51% of published novelists agreed that AI is likely to displace their work. That is an expectation, not a measurement — writers forecasting their own obsolescence. It is a striking vote of no confidence from the people closest to the work, and it colored much of the fair's mood.
For more context, read Amazon KDP's April 2026 Rule Changes: What Indie Authors Should Check.
For more context, read aap statshot q1 2026.
For more context, read Stuff Your Kindle Day Returns on August 10, 2026.