So. AI.

(Or… but haven’t we already seen how this will play out in publishing?)

I couldn’t help but be dismayed when I read in Publisher’s Lunch that the CEO of Bertelsmann (the company that owns Penguin Random House) is excited about the possibilities of AI generated content in publishing in the Financial Times.

Overall, [Rabe] sees the technology as ‘on balance . . . probably more of an opportunity.’ For example, ‘He said there were even opportunities for authors to augment their output by feeding software with their previous output to generate new content.’ Rabe added, ‘If it’s your content, for which you own the copyright, and then you use it to train the software, you can in theory generate content like never before,’ he said. Plus, ‘Rabe also argued that a proliferation of AI-generated books and articles of varying quality would increase the value of works from respected sources.’
— Publisher’s Lunch on “AI is an opportunity for creative industries, says Bertelsmann boss,” Financial Times.

It reminds me of the indie publishing wave about fifteen years ago.

I was initially very excited when the Amazon team was going in-person to literary agencies in the late ‘oughts to persuade agents to grant ebook rights to publishers so that there was content for Amazon’s Kindle. While they courted traditional publishers and agents, they also launched KDP in November 2007 - a platform for authors to independently publish their books for sale and reading on the Kindle.

In 2011, when indie ebooks were gaining traction and I had the time to devote to writing, promoting, and learning from others in the indie ebook community, I published a dictionary of publishing terms, a YA short story, and the next year a middle grade novel.

The KDP platform was three years old, but now I wonder if even then, I was already too late…

The people who got onto Amazon’s (KDP), B&N’s (Nook) and Smashwords self-publishing platforms first and with polished content ready to go had a shot at doing tremendously well. In 2012, NPR, Entertainment Weekly, etc were all talking about runaway indie bestsellers - e.g., the Trylle trilogy’s Amanda Hocking, Wool’s Hugh Howey, The Martian’s Andy Weir.

Being a high-profile indie bestseller was great for the authors…and really, really great for attracting new writers to produce Kindle content.

But then Amazon’s algorithms began to reward creators who published new content every 30 days, and the people that had the best chance of doing well were now those that could produce a lot of material very quickly.

Amazon also price matched free ebooks (and then eventually just let you set your ebook to free for a period of time if the book was exclusively on Kindle) and a whole industry of bloggers and social media accounts that promoted free ebooks was born.

In that system, authors (and eventually publishers) were paying bloggers to advertise free ebooks to guarantee that their free ebook promotion was noticed, in the hopes that then someone would check out and pay for the rest of the series.

I indie published in 2011/2012, and then took a break for grad school. When I came back around 2015, it was clear that the system had changed. I knew that I did not have the ability to produce enough content that I was comfortable publishing in a time frame that would make it visible to Amazon’s algorithms.

Readers were also exhausted. Like many others, I had found myself with a Kindle stuffed full of books of varying quality as traditional publishers, earnest indie publishers and scammers uploading public domain, pirated, and/or computer generated gibberish began to compete in the same system.

This brings me back to Rabe’s vision for AI in publishing - which sounds to me like a way for a publisher to try to flood the market with works by their best-selling authors.

But scammers have done this on KDP for years now - because the Amazon algorithm is designed to incentivize a massive quantity of work in as short a time frame as possible with no incentive for better quality.

And the winner won’t be the author, or the publisher. It will be Amazon.

In talking about scammers inflating the page counts of their books to game the Kindle Unlimited payout system, Gaughran says:

…it hurts authors much more than Amazon. They might see it as only affecting 0.2% of books or whatever, but the top scammers are making over $100,000 a month – money that comes from the author fund, not Amazon’s end. These people gaming the system will roll a huge chunk of that back into advertising too, which either brings readers to the website, or goes directly back into Amazon’s pocket via Amazon Ads.
— David Gaughran, Kindle Author “Plagiarism, ‘book-stuffing’, clickfarms...the rotten side of self-publishing,” Guardian.com

Like Gaughran mentioned in the Guardian article, more and more of publishing’s money is cycling back to/through Amazon.

Including my own.

I have done Amazon advertising for my traditionally published picture books recently, and it did boost sales and lead to more reviews.

However…when I look at how much I paid to Amazon vs how many books I sold in response (and as a traditionally published author, what my royalties are on those books), the winner was - and was always going to be - Amazon.

The first authors to jump on Amazon ads for their books probably are the ones that we’re all still seeing on repeat now across the platform, with thousands of reviews and not as much financial investment needed to keep the cycle going because of the initial boost. They are the proverbial teddy bears that Corey Doctorow talks about in this article on the ‘enshittification’ of social platforms - the early adopters who were boosted so that other users would be drawn in.

Much like the success of the early adopters of KDP brought in waves of content creators for the Kindle, these books with thousands of reviews on Amazon work to bring advertising dollars to Amazon as we (as creators and publishers) hope to replicate the same success and be seen in the onslaught of noise.

We cannot all - authors, illustrators and publishers - ignore Amazon as a distributor for a number of legal, financial and human nature-type reasons. As an agent and as an author, I want publishers to put marketing dollars towards my books in the places that will result in sales. Even as I type all this.

I don’t have answers, but I do have a lot of concerns:

‘[Rabe] said there were even opportunities for authors to augment their output by feeding software with their previous output to generate new content.’ Rabe added, ‘If it’s your content, for which you own the copyright, and then you use it to train the software, you can in theory generate content like never before’
  • Will authors be paid at the same rate for work generated using AI? Or will this be an excuse to lower advances, since the publisher will argue that creators can now produce more work in the same period of time?

  • Will authors retain the right to be the only one who can do this with their work? [See the SAG-AFTRA concern about studios licensing an actor’s likeness indefinitely.]

  • How many times can an original piece of creative work be fed through an LLM and spat out before it is no longer considered an original work? (E.g if I drew ten pictures of coffee cans, fed them into AI, asked them to draw me a coffee can, then fed that into the AI and asked them to draw me a coffee can… when - or does - it ever stop being considered my original work? In this cycle, if 10 coffee can images were human generated and the next 200 were AI generated, is the 211th coffee can image still copyrightable by the human that drew the first 10?)

  • How will authors be compensated when their work is fed into software for LLMs? [Litigation is already ongoing - but when LLMs are ubiquitous, how would one even know that their material was scraped to bring suit? This article will likely be fed into an LLM as part of a large batch of material at some point in the future.]

  • How can an author keep up with the promotion and marketing of so many extra titles per year, should AI be used as Rabe proposes?

  • Will the publisher be hiring more editorial, art, marketing, publicity and sales staff to support all these additional AI-generated/AI-assisted books?

  • How can we protect illustrators when AI is good enough to illustrate a picture book? Will publishers commit to hiring human illustrators for their books? Can agents push for this in contracts?

a proliferation of AI-generated books and articles of varying quality would increase the value of works from respected sources.
  • Haven’t publishers, media companies and news networks been seeing the exact opposite? [Stephen T. Colbert introduced us to truthiness almost ten years ago...]

  • Wouldn’t a proliferation of AI-generated books and articles of varying quality just be a sanctioned version of the KDP scammers glut of content to game the system?

  • How does the publisher intend to get their AI-generated books and articles seen above the noise of everybody else’s AI-generated books and articles? Especially AI-generated content at Amazon’s own publishing imprints?

  • How will readers find extra time for all this content? Reading time is limited by the number of free hours in a day. Won’t the same number of reading hours just be divided among a larger sea of authors/books…. making it harder for any one book or author to earn out their advance?

  • Who will buy all these AI-generated books if we aren’t paying human content creators for their work?

A majority of all books and ebooks in the US are bought on Amazon. And I can’t help but look at the lessons of indie ebook publishing and worry that the big winner of AI generated content won’t be the authors, illustrators, editors or even the publishers.

It’ll be Amazon.

So I hope that the major publishers will be standing with their authors, illustrators and staff when it comes to AI/LLMs in the industry. Because from where I stand, the biggest threat to their financial bottom lines is not the cost of paying their creatives and staff. It’s a system where Amazon doesn’t need them at all.

Previous
Previous

Try Three Times and Fail: A quick breakdown of traditional picture book structure

Next
Next

Thinking Like a Literary Agent