How AI-Generated Books Could Hurt Self-Publishing Authors

This content has been archived. It may no longer be accurate or relevant.

From Jane Friedman:

Just two days after the Maui wildfires began, on Aug. 10, a new book was self-published, Fire and Fury: The Story of the 2023 Maui Fire and its Implications for Climate Change by “Dr. Miles Stones” (no such person seems to exist). I learned about the book from this Forbes article, but by then, Amazon had removed the book from sale. Amazon had no comment for Forbes on the situation.

Curious about how far the book might have spread, I did a Google search for the book’s ISBN number (9798856899343). To my surprise, I saw the book was also for sale at Bookshop and Barnes & Noble. I tweeted about the situation, noting that IngramSpark, a division of Ingram, must be distributing these books to the broader retail market. My assumption was that retailers, in particular, Bookshop, would not accept self-published books coming out of Amazon’s KDP. (Amazon KDP authors can choose to enable Amazon’s Expanded Distribution at no cost, to reach retail markets outside of Amazon.)

It turns out my assumption was wrong. Bookshop does accept self-published books distributed by Amazon, and here things get a little convoluted. Amazon Expanded Distribution uses Ingram to distribute; Ingram is the biggest book distributor and there isn’t really any other service to use for distribution as far as the US/UK.

However, Bookshop’s policy is not to sell AI-generated books unless they are clearly labeled as such, so Fire and Fury was removed from sale after they were alerted to its presence. Bookshop’s founder Andy Hunter tweeted: “We will pull them from @Bookshop_Org when we find them, but it’s always going to be a challenge to support self-published authors while trying to NOT support AI fakes.”

And now we come to why self-publishing authors have reason to be seriously concerned about the rising tide of AI-generated books.

  • Amazon KDP is unlikely to ever prohibit AI-generated content. Even if it did create such a policy, there are no surefire detection methods for AI-generated material today.
  • Amazon KDP authors can easily enable expanded distribution to the broader retail market at no cost to them. It’s basically a checkbox.
  • Amazon uses Ingram to distribute, and Ingram reaches everyone who matters—bookstores, libraries, and all kinds of retailers. Ingram does have a policy, however, that they may not accept “books created using artificial intelligence or automated processes.”
  • Based on what happened with Fire and Fury, Amazon’s expanded distribution can make a book available for sale at Barnes & Noble and Bookshop in a matter of days.

If the rising tide of AI-generated material keeps producing such questionable books—along with embarrassing and unwanted publicity—one has to ask if Barnes & Noble and Bookshop might decide to stop accepting self-published books altogether from Ingram or otherwise limit their acceptance. Obviously not good news for self-published authors, or Ingram either.

What are some potential remedies?

  • Ingram is an important waypoint here. They’ve put stronger quality control measures in place before. Perhaps they can be strengthened to prevent the worst material from reaching the market outside of Amazon.
  • Amazon’s Expanded Distribution requires that authors use Amazon’s free ISBNs. Would it be possible for retailers to block any title with an Amazon ISBN? (ISBNs identify the publisher or where the material originated from.) While that may be unfair to honest people who prefer to use Amazon’s Expanded Distribution, such authors/publishers would still have the option of setting up their own IngramSpark account. IngramSpark has no upfront fees and also provides free ISBNs.
  • Maybe IngramSpark or other retailers put a delay on making Amazon’s Expanded Distribution titles available for sale. Amazon already states it can take up to eight weeks for the book to go on sale. So why not make such titles wait?

Free ISBNs unfortunately contribute to this problem

ISBNs are a basic requirement to sell a print book through retail channels today. In the US, it is expensive to purchase ISBNs—it’s nearly $300 for ten. Amazon KDP does not require authors to purchase ISBNs and will give you ISBNs for free all day if you need them. Over time, others like IngramSpark and Draft2Digital have also made ISBNs free to make it easier for self-publishing authors to distribute their work.

While it’s admirable to lower the barriers for authors who have limited funds, free ISBNs are supercharging the distribution of AI-generated materials to the wider retail market. An immediate way to stem this tide of garbage in the US market? Stop giving out free ISBNs. Make authors purchase their own.

There’s a huge advantage to making authors purchase their own ISBNs: it creates an identifiable publisher of record with Bowker (the ISBN-issuing agency in the United States). The publisher of record would be listed at retailers. Currently, fraudsters using Amazon KDP are able to hide behind Amazon-owned ISBNs; their books are simply listed as “independently published.” It would be marvelous to take away that fig leaf. Sure, fraudsters could create sham entities that mean nothing and are unfindable in the end, but at least you could connect the dots on all the titles they’re releasing—plus Bowker would see who’s doing the purchasing and possibly put their own guardrails in place. My hope is these entities would choose not to buy ISBNs at all and this activity would become limited to the backwaters of Amazon.

Link to the rest at Jane Friedman

PG has the greatest respect for Jane and has been appreciating her blog posts for years, but ISBNs are nineteen-sixties technology. Any industry other than publishers\booksellers would have ditched it for something more effective, useful, and easier to administer decades ago.

7 thoughts on “How AI-Generated Books Could Hurt Self-Publishing Authors”

  1. One wonders when commercial publishers will go after Bowker for its abuse of a monopoly position not granted by a government actor with the authority to grant a monopoly. Oh, wait a minute, the commercial publishers have since the 1960s successfully used the particular pricing scale offered by Bowker to establish an entry barrier against small publishers — a pricing system that did not reflect any actual “economies of scale” then or now, but instead appears… collusive.

    Of course, “reg’lar readers” don’t have antitrust standing against Bowker, and would have antitrust standing against the commercial publishers only to the extent that those “reg’lar readers” could prove that they were paying a higher price for the goods actually offered due to these pricing practices.† And intermediaries who are not the true end users (there’s utterly implausible, but utterly fossilized, case law that “obtaining a third-party compliance certification” is an intermediate transaction outside the scope of the Sherman and Clayton Acts absent proof of either actual collusion or systematic favoritism in the grant/denial of certification), so smaller publishers who in the past were unable to get their stuff into bookstores without an ISBN can’t sue either, even assuming that the statute of limitations isn’t long gone.

    † Under the dubious Chicago School reasoning that took over antitrust doctrine starting in the early 1970s, mere “restriction of choice” does not qualify for antitrust standing — consumers and other end-users must couple any harms from restriction of choice to a higher price paid or an objectively-verifiable lower quality received. Objectively-verifiable lower quality received… literary merit… yeah, that’s something judges can agree on (because standing is considered, wrongly, a pure question of law, so it never reaches a jury).

    • Do you really expect a populist consumer society to protect producers/vendors at the expense of consumers?

      Maybe in a Brane where 4 of the last 5 administrations weren’t populists (the 5th got elected as one but couldn’t follow through because of an incident) but in this one, “vote your pocketbook” and “everybody gets $2000 if we win” sets the real rules, not theoreticians.
      (With worse to come.)

  2. Respecting your opinion, PG, but I tend to agree with Dean Wesley Smith, that Jane Friedman is an apologist for traditional publishing, and a fairly transparent one.

    Know what else could hurt self-publishing authors? Banning or otherwise getting rid of free ISBNs, which list the publisher of record as whichever company doles out the ISBN. Throwing out the baby with the bath water, that.

    I suspect Jane’s be-all end-all is the elimination of self-publishing altogether. The real problem here is how self-published authors and their reasonably priced ebooks are harming traditional publishing and their excessively priced ebooks.

    • Yes, JF’s must have been asleep for the past 20 years: “…Ingram reaches everyone who matters—bookstores, libraries, and all kinds of retailers. ” Consider where the majority of book sales take place, she’s willfully blind to the 800 lb gorilla in the room. And that TradPub is certainly no guarantee of quality.

  3. As one of those durned foreigners, I always thought it silly that Bowker charged for the hallowed ISBN. It’s a costly tollgate in front of the indie publisher. Now, I’m thinking Jane wants Bowker et al (et al soon to be eliminated, apparently by Jane) to charge an even more ridiculous fee. Talk about being knee-deep in gatekeepers, and she would like another one. How about making it a requirement for all American authors to have an agent, a properly vetted editor, and only a publisher of record? That would solve the perceived problems immediately.

Comments are closed.