The CEO of Barnes & Noble believes that AI-generated books are acceptable. He is mistaken.

      James Daunt, the CEO of Barnes & Noble, recently spoke with NBC News, and he made a statement that has lingered in my thoughts. When queried about books written by AI, Daunt responded, “Yes, I have actually no problem selling any book, as long as it doesn’t masquerade or pretend to be something that it isn’t, and that it has an essential quality to it, and that the customer, the reader, wants it.”

      At first glance, this appears quite reasonable. If readers are able to clearly see the labeling, they can make an informed choice. However, upon further reflection, this perspective raises several significant questions.

      Is simply labeling them truly sufficient?

      Barnes & Noble ranks as one of the most influential retailers in the publishing industry. When the largest book retailer in the U.S. indicates that it is open to AI-generated books, it communicates to publishers, agents, and authors that this is a valid category of products.

      Consider the significance of a real book. An author dedicates months, sometimes years, to research, writing, revising, and creating a cohesive work, while infusing it with their unique experiences. This human aspect is what differentiates books, prompting readers to explore various authors' takes on the same subject.

      In contrast, AI takes the knowledge it has gleaned from human experiences, removes the human element, and produces something devoid of that essence. While it may achieve grammatical precision, structural excellence, and even an engaging narrative, it lacks the human touch that gives books their distinctive quality. At most, it can simulate that touch, drawing from the wisdom embedded in books authored by humans.

      When a major retailer simply accepts AI books as long as they come with a label, it undermines the understanding of a book as a product of human creativity. Moreover, who determines the criteria for labeling what constitutes an AI-written book? Is a subtle label on an obscure page sufficient, only discoverable to those actively seeking it?

      Even if there is a clear label, does it really matter? Would you allow a thief into your home just because they have a label indicating their intentions? The analogy is absurd. Any AI-generated book, regardless of its quality, is akin to a thief in disguise, having appropriated narratives from human-written works without permission.

      The human cost of allowing AI books in our bookstores

      Every bookstore has finite space. Permitting AI books into our stores means that each one displaces a book written by a human. Without an effective system to distinguish between the two, which Barnes & Noble appears to lack, it would be challenging for readers to tell them apart.

      Daunt even conceded that Barnes & Noble may unknowingly be selling AI-generated books. “We have 300,000 titles across all of our stores. Do we think that some of those may be AI? The chances are that they are, but we’re not really conscious of them,” he stated in the NBC News interview. This is hardly the reassuring revelation he believes it is.

      What you see is what you purchase. If numerous readers enter the store and notice AI books prominently displayed, some will inevitably pick them up. This could result in profits for a large corporation or an AI entrepreneur treating books as a new venture. That represents a sale that could have gone to a deserving author.

      I’m not suggesting that all human-written books are exceptional. I’ve penned some mediocre ones myself. However, even if a book is poor or not to your liking, it’s evident that real effort went into it, so the financial impact isn’t as harsh.

      Consider how you would feel if your books were written by a prompt. Moreover, given that AI can produce books considerably faster than humans, the market will become inundated with these publications if we allow them in. The e-book market is already saturated with AI-generated content; we certainly wouldn’t want our bookstores to reflect that.

      This is not occurring in isolation

      It would be one thing if Barnes & Noble were making this decision alone. However, it’s part of a broader and troubling trend.

      Vox Media and The Atlantic both entered into agreements with OpenAI, permitting the company to use their entire content archives for training its models. The New York Times has signed its first AI content licensing deal with Amazon, and similar multi-year deals have been made by USA Today, Condé Nast, and Hearst with Amazon.

      AI licensing agreements are rapidly becoming a significant revenue stream for publishers. Thus, publishers receive compensation, making these arrangements seem justified. Meanwhile, the writers whose work is utilized to train these models are largely compensated.

      The pattern is unmistakable. Initially, media companies license their content to AI. Following that, AI generates new content from the licensed material, which retailers then sell. This cycle will continue until human authors are obsolete, leaving us with a flood of AI-generated content and reflecting on how we arrived at this point.

      Books represent one of the last areas where human creativity is not