Waterstones Would Sell AI Books—But Nobody There Wants To

Waterstones CEO James Daunt just gave the most honest answer yet about how retail will handle AI-generated content: they'd sell it if customers demanded it, but personally, he doesn't think that'll happen.

"There's a huge proliferation of AI-generated content and most of it are not books that we should be selling," Daunt told the BBC's Big Boss Interview podcast. But he added the decision would ultimately be "up to the reader."

The 62-year-old bookselling veteran—who's heading into his 36th Christmas season in the trade—was more direct about his staff's likely reaction: "Do I think that our booksellers are likely to put those kind of books front and centre? I would be surprised. We as booksellers would certainly naturally and instinctively disdain it."

This tension—between retail neutrality and human judgment—captures exactly how businesses will navigate AI content as it proliferates across creative industries.

The "Clearly Labeled" Fantasy

Daunt's condition for stocking AI books is transparency: "As long as it doesn't pretend to something that it isn't." Any AI-generated book would be "clearly labelled as such," he emphasized.

But this assumes AI labeling remains straightforward. What about books where AI handled research but humans wrote prose? Or humans outlined plot while AI generated dialogue? Or extensive editing of AI drafts? The boundaries between "AI-generated" and "human-authored with AI assistance" aren't binary. They're gradient.

Publishers already struggle with this classification problem. Many authors use AI for research, outlining, even first drafts that get extensively revised. At what percentage of AI contribution does a book become "AI-generated"? Who decides? Who verifies?

Waterstones can mandate disclosure, but enforcing it requires trust that publishers accurately report AI involvement. That trust seems optimistic given economic incentives to obscure synthetic content's role.

Why Readers Might Actually Want This

Daunt acknowledges that technology firms "are spending trillions and trillions on AI and maybe it's going to produce the next War and Peace. And if people want to read that book, AI-generated or not, we will be selling it."

This conditional is key. He's betting readers value "a connection with the author that does require a real person." But what if that assumption is wrong? What if significant audiences don't care about authorial connection as long as the story entertains?

Genre fiction already demonstrates this pattern. Many readers consume romance, thrillers, or fantasy primarily for plot and pacing rather than authorial voice or literary craft. If AI generates competent genre fiction faster and cheaper than humans, some readers will choose it—particularly for disposable entertainment consumed once and forgotten.

The economics favor this outcome. AI-generated books cost almost nothing to produce and can flood markets with content calibrated to reader preferences through data analysis. Human authors can't compete on volume or price. They can only compete on quality dimensions that readers actually value—and not all readers value the same things.

The Bookseller Judgment Filter

Waterstones' actual defense against AI content isn't policy—it's culture. Daunt rebuilt the company by empowering store managers to curate inventory for their communities. Staff write recommendations, choose books of the month, control display tables.

This decentralized curation creates natural resistance to AI content. Individual booksellers who "instinctively disdain" synthetic books won't prominently feature them even if corporate policy permits stocking. The human judgment layer filters AI content through taste and professional pride.

But this only works while booksellers maintain that authority. If customer demand for AI books grows significant enough to impact revenue, corporate pressure to stock and promote them increases. Daunt's current position—"we'd sell them if readers want them"—contains the seeds of that reversal.

What This Reveals About Retail's AI Dilemma

Waterstones' position previews how retailers across categories will handle synthetic content: initial resistance based on quality standards and brand identity, gradual accommodation if customer demand materializes, eventual normalization once AI content becomes indistinguishable from human work.

The "clearly labeled" requirement provides temporary comfort but won't survive contact with messy reality. Labels erode as collaboration between humans and AI becomes standard practice. Quality filters weaken as AI capabilities improve. Professional disdain fades as economic pressure mounts.

A Cambridge University report found that more than half of published authors fear AI replacement, with two-thirds saying their work trained AI models without permission or payment. These aren't hypothetical concerns. They're describing current reality that retail hasn't yet fully confronted.

The Question Retailers Won't Answer

Daunt's position is admirably honest about the contradiction: Waterstones would sell AI books but hopes customers won't want them. This avoids the harder question: what happens when AI-generated content becomes good enough that most readers can't tell the difference?

At that point, "instinctive disdain" from booksellers becomes professional snobbery rather than quality curation. The connection to a "real person" that Daunt believes readers value becomes nostalgia rather than necessity. And retail neutrality—selling what customers want—means stocking whatever content moves regardless of its origins.

For businesses navigating AI content in their industries, Waterstones offers a preview: you'll start with principles about quality and authenticity, make exceptions for customer demand, and eventually accommodate synthetic content while hoping nobody notices the compromise. At Winsome Marketing, we help companies develop positioning strategies that account for synthetic content's inevitability rather than pretending human judgment will permanently prevail. Sometimes honesty means admitting you'll sell what customers buy—even if you hope they won't.