Barnes & Noble's CEO, James Daunt, went on the Today show and said he had "no problem selling any book," AI-written ones included, "as long as it doesn't masquerade or pretend to be something that it isn't." The internet read that as a green light for machine-made books, and within days thousands were calling to boycott the chain. Daunt walked it back: B&N doesn't knowingly sell AI books and demands publishers label them — "a straightforward rejection of AI books."

1. Label It, Don't Ban It (James Daunt / B&N)

A bookseller shouldn't be the one deciding what counts as a real book.

A bookstore's job is to stock what readers want, not to police how a book got written. Daunt's line is that the honest fix is disclosure, not prohibition — if a book says it's AI-written and isn't ripping anyone off, the reader can decide for themselves.

Banning any book is bad. Where would the line even sit — a book that's 100% machine-written, or half? That call belongs to publishers, not the cash register, and no serious publisher is putting out AI novels anyway. He calls the whole thing common sense.

2. This Is a Sellout (authors, the Authors Guild)

AI books are trained on our work and bury us on the shelf — a bookstore shouldn't touch them.

AI books are built out of stolen writing, and stocking them rewards the theft. Authors say the machines were trained on their work without consent or payment, then churn out cheap imitations that crowd real books off the shelf.

The backlash was immediate and personal. Writer and former B&N employee Kathlin Finn said she'd stop shopping at and promoting the chain until it changed course — one of thousands who called for a boycott.

Writers are policing this themselves because no one else will. The Authors Guild has handed a "Human Authored" seal to more than 3,000 authors across 5,000-plus titles, and CEO Mary Rasenberger calls it the next best thing to the mandatory AI labels that don't yet exist.

3. The Labels Are Meaningless (skeptics, Originality.AI)

You can't tag what you can't detect, and the slop is already inside the catalog.

You can't label what no one can reliably identify. Even the Authors Guild admits there's no dependable way to detect AI writing — its own seal vouches for a process, not for the actual text on the page.

The slop isn't coming; it's here. The detection firm Originality.AI found that 82% of the herbal-remedies paperbacks Amazon published in 2025, and 77% of the "success" self-help titles it analyzed, were likely AI-written — often under invented author names.

If a store genuinely can't tell, a pledge not to knowingly stock AI books doesn't mean anything. The real fight is upstream — at the platforms flooding the market and the training data behind the machines — not at the Barnes & Noble register.

Where This Lands

Daunt's stance is either principled — a bookstore has no business ruling on what counts as a "real" book — or a tidy way to avoid picking a side. Authors see something simpler: their work scraped and resold, with the chain that's supposed to be on their side shrugging. And the skeptics keep pointing back at the catalog, where the machine-written books are already shelved and mostly undetectable. Nobody in this fight can reliably tell a human book from a robot one. Until someone can, the argument over labels and bans is really an argument about who's responsible — the store, the publisher, or the platforms that let the flood in.

Sources