James Daunt didn’t set out to make news with a rule about book covers. In a live Today show interview with Jenna Bush Hager in May 2026, when asked if Barnes & Noble would stock a book written by artificial intelligence, the chain’s CEO gave an answer that sounded like the standard for direct disclosure. His reaction over the course of days suggested that most people listening were hearing something else.
What Daunt actually said
Daunt’s full answer, This was reported by Today.comwas: “I don’t really have a problem with selling any book, as long as it’s not masquerading or pretending to be something it’s not, and it has an underlying quality and the customer, the reader, wants that… As long as an AI book says it’s an AI book and doesn’t pretend to be something else, as long as it’s clearly stated and the customer wants to buy it, we’ll stock them.” Barnes & Noble carries about 300,000 titles, he added This was reported by NBC Newsand some of them have already been written by AI without the company’s knowledge, a detail that undermines the neatness of the disclosure standard it just proposed. If a company can’t tell if the books currently in its inventory were written by AI, the rule requiring future disclosure does nothing about the books already on the shelf.
A statement of reaction was not expected
Fortune reported about it Daunt’s comments “quickly sparked outrage on social media, including plans to boycott the company until it changed its AI policy.” The outrage wasn’t really an argument over tagging mechanics, whether the sticker should be on the front or the back, or whether the words should say “AI-assisted” or “AI-written.” It was a rejection of the idea that a book written by artificial intelligence should generally be on the same shelf as a book written by a human, provided the packaging is honest about which is which. Daunt’s standard viewed AI authorship as an advantage to be disclosed, such as spice level or runtime.
A significant portion of the backlash saw it as a flaw that should be kept out of the store altogether. Specific reactions circulating in the days after the interview focused on calls for a complete boycott of the retailer rather than calls for a clearer labeling policy. People outraged by its existence are asking for its removal.
Why disclosure is always the first suggestion
Executives in Daunt’s position pursue disclosure requirements before public dismissal for a predictable reason: while promising disclosure is procedurally simple and difficult to enforce, promising dismissal is procedurally difficult and relatively straightforward to verify after acceptance. Telling a supplier to “label” costs the retailer next to nothing.
Telling a supplier that “we won’t carry it at all” requires the retailer to actually be able to tell the difference between thousands of titles from publishers and hundreds of thousands of self-published authors at the point of receipt, a more difficult operational challenge than printing a sticker.
Daunt’s original answer is optimized for things that are easy to handle in a live interview. His revised answer, days later, acknowledged that the easy version would satisfy no one.
What disclosure cannot do
This is the gap that occurs when disclosure-based standards are proposed as a solution to a disagreement of values rather than an information gap. The tag works when the objection is “I didn’t know” – tell the customer and the transaction is reported. The “I don’t want this sold here regardless of what I know” objection, which is closer to what boycott threats actually mean, doesn’t work.
No clearer statement resolves the disagreement over whether the product should be on the shelf in the first place.
Neither answer solves the scale problem
Even the exception standard Daunt described in Fortune , active measures to keep AI-generated books out of the catalog, reaches the same practical limits he noted in the original interview: 300,000 books in circulation and no reliable way to identify AI-generated prose.
This is a narrower gap than the gap between disclosure and exclusion, but it is the same form of the problem: a retailer may adhere to a policy without being able to fully verify that the policy is being followed, for the same scale of disclosure that led the Society of Authors to agree to an honor code rather than a manuscript review process for its AI analysis.
Daunt backed off his answer
Apparently, Daunt noted the difference between what he said and what people heard. A few days after the original interview, he told Fortune directly: “Our position is that we do not sell AI books, as far as we know; we take proactive measures to remove all AI-generated books from our online catalog and never knowingly backorder them in our stores… This is a direct rejection of AI books.” This is a different policy than the one he describes in Today. The first version was the disclosure standard: label it and we’ll sell it. The second is the standard of exception: we try not to sell it at all. The gap between these two announcements, made by the same executive about a week apart, is a fairly accurate measure of how quickly the disclosure framework collapsed under public backlash.
When threatening to boycott a bookstore over its disclosure policy, what readers want could never be better served by disclosure. They wanted the books to go, not labeled. Daunt’s revised response suggests that he understands this much more quickly than his original framework.
The episode is a useful pointer for anyone watching how retailers and publishers generally respond to questions about AI: the first public response is procedurally convenient, and the real position emerges only after meeting an audience that already knows what the comfortable response actually wants.
