Frankfurt opens today. It’s the biggest book fair on the planet, the 78th one, and it runs through Sunday. At yesterday’s opening press conference, Sebastian Guggolz, who chairs the German publishers and booksellers association, used a line I can’t shake. He called AI a double raid. First the books, then the thinking.
The books part we know. Companies trained on millions of copyrighted works without asking and without paying. So the association put out what it calls the Frankfurt Appeal, aimed at the EU Commission, the EU Parliament and the German government. It asks for transparency about which works went into training, fair licensing, and copyright you can actually enforce. 120 publishers have already signed. Fair enough. Overdue, honestly.
The second half is what got me. Guggolz pointed to the latest PISA study, which found a third of 15-year-olds can’t read well enough. He also said the number of book buyers aged 10 to 15 dropped 30.6% in 2025 compared with the year before. Then he said this (my translation from the German): if you can’t read for meaning, you can’t check an AI’s answer at all. All you can do is believe it.
Surprise. The machine that answers everything needs a reader who can tell when it’s wrong. And we’re turning out fewer of those readers.
Juergen Boos, the fair’s outgoing director, summed up the industry’s split in one breath. How does publishing use AI to get better, and how does it protect its works? Both questions, at the same time. He said that tension is the backbone of this year’s program. Been there. Every writer I know lives in that sentence.
Now the little world. Less AI, more me.
Circle is being built on reading. Not detecting, not scoring. Reading. A member requests an assignment, sits with somebody else’s chapter or video, and then submits a structured and prompted review on site: what works, what doesn’t, where they got lost, and why. That review is the credit you earn toward getting your own work reviewed, or spending elsewhere on site. You don’t get in line by posting. You get in line by paying attention to someone else’s work.
I didn’t build it that way because of PISA scores. I built it that way because that’s how writers get better. But reading Guggolz yesterday, it hit different. Close reading is the skill holding the whole thing up. It’s how you catch a weak scene. It’s also how you catch an AI answer that sounds right and isn’t. Same muscle.
It’s also why Circle encourages disclosure by role instead of detector gates. Honesty about using AI is the backbone. Every member that contributes knows and understands the responsibility to be honest in representing their work. Everyone knows they pledged this honesty by the simple fact that their work appears here.
So the Frankfurt Appeal asks AI companies to show their sources. Good. I’d sign it. But the other half of the fix doesn’t come out of Brussels or Berlin. It comes from people who still read slowly and carefully, with a pencil in hand. A workshop is where that habit gets practiced on purpose, every week, on real work.
GAPOZ Press is the people’s publisher. The people still have to be able to read. That part’s on us.
Also on the beat
- The Society of Professional Journalists approved the sixth revision of its 100-year-old Code of Ethics, which now addresses AI and machine-generated content, with 93% of members voting yes. SPJ, Oct 4
- In Korea, “Narrator: AI” labels are showing up on popular audiobooks, since AI can bring production costs down from about $3,750 a title to under $750, and the Korean Publishers Association is funding 58 AI audiobook titles. Kyunghyang Shinmun, Oct 7
- Publishers finally have AI companies buying their content, but the buyer still sets the price. The Media Copilot, Oct 6
- Regina Brooks, president of the Association of American Literary Agents, and Originality.AI’s Jon Gillham talk about AI in book publishing and how detectors work. ABC News, Oct 6
- Random House bought Kim Johnson’s YA thriller Lucky No. 8, about a teen competing under AI surveillance to get into a university founded by tech moguls. Publishers Weekly, Oct 6
How this was made
AI assisted research and a first draft of this Circle View. A human reviewed, edited, and is accountable for the judgment and the final take. Disclosure by role — not undetectable.
- A score is not a confession → Page 1 Still think a percentage is proof? Read Do AI detectors work on fiction?
- A Detector is NOT a Judge → Page 1 The short version of why a score can’t judge a book: Do AI detectors work on fiction?
- Prove you didn’t use a machine → Page 1 If someone’s asking you to prove a negative, start here: Do AI detectors work on fiction?
- Disclosure that never faces the reader → Page 2 Here’s how we keep it simple in Circle, no formulas: How Circle handles AI
- How did you use AI? → Page 2 We ask once, and your word does the rest: How Circle handles AI
- Put it on the Imprint Page → Page 2 Why I’d rather skip the label tiers altogether: How Circle handles AI
- Watermarks don’t sign your name → Page 1 Watermarks, detectors, same problem: Do AI detectors work on fiction?
- They knew the shelf would get crowded → Page 2 Slop is the problem, not the tool. Here’s where Circle stands: How Circle handles AI
- Label the voice → Page 2 How we handle AI in Circle, without the labels: How Circle handles AI
- Byline is no Costume → Page 2 Your name, your word, your work. That’s the deal in Circle: How Circle handles AI
