A lot has changed since ChatGPT was released in November of 2022, and a lot more change is ahead of us in the publishing industry whether we like it or not. How you may feel about the creation of AI aside, it’s here to stay, it’s devouring water and energy at a ferocious pace, and publishing CEOs have been saying the quiet part out loud for years already. There are important things authors need to know about what we’re seeing so far and what may be on the horizon.
If you’re not sure if your book was used to train AI, you can search for it in The Atlantic’s search tool covering the 7.5 million books in the LibGen data set.
I’ll address the elephant in the room. If your book was stolen to create Large Language Models (LLMs), I’m really sorry that happened to you. It’s piracy, pure and simple, and never should have happened. For those of you who think, what’s the big deal? I encourage you to learn about what happened to copyright and licensing revenue for educational books in Canada and maybe you’ll change your mind. Systems collapse when IP is stolen. It’s economics. It’s why Open Access is such a tricky topic, but we’ll cover that another time.
Authors are already deeply under-compensated for their time and work, so I say, go get yours. For more information, visit the Authors Guild.
Quietly, and almost everywhere at once. AI is already built into the tools marketers and publicists use every day, agents are on the lookout for AI-generated submissions, and our media colleagues are buried under AI-generated pitches.
Marketers are already using tools that embed AI into programs like Canva, Photoshop, Illustrator, social listening tools, and more. Publicists can utilize AI integrated into their media databases, and I can’t seem to get AI podcast pitching platforms to stop trying to sell to me.
Literary agents are using their brains to tell them if query letters and chapters are AI generated, and more recently Pangram. And yes, some authors are trying it: The New York Times reported earlier this year that authors are using AI to write entire books. This is my warning to not submit AI-generated books to agents. They do not appreciate it. If they wanted AI-created books, they have the talent and skills to effectively create them themselves. Why would they take a tiny slice of your advance for all the work they do, when they could spin up something very likely better on their own?
In public relations, we’re already seeing some problems with AI. Take those podcast platforms I mentioned earlier. They’ll build lists, write pitches, and blast those emails out, all with barely a word from a publicist. Our media colleagues were already fielding several hundred emails each day before the advent of GPTs; now it’s thousands. How do they wade through all of that? The answer is simple: they can’t. I’m sure plenty goes to spam, they will start to blacklist certain domains, and we’ll carry on as we always do.
Email is an important tool for us, but it was always just a tool. Relationships are what have allowed us to achieve what we do for our clients, not spamming as many journalists as we can in one shot. That’s never been our method at Page One. That said, we would be foolish to not review the AI recommendations in Muck Rack when we’re building a list. Sometimes it suggests someone we didn’t think of or someone we just plain forgot. It happens and I’ll take the occasional reminder, but I can also waste 30 minutes refreshing those recommendations and never getting anything good—what a waste of time. Nine out of 10 of the journalists it recommends aren’t a fit, and we’re definitely not adding people to media lists just for the hell of it.
Yes, and more than you might think. Publishers are using AI behind the scenes for forecasting and operations, out in the open for jacket design and social media, and, in some departments uploading the manuscripts themselves.
If the history of the last 25 years in publishing tells us anything, it’s that publishers are always looking for ways to save money. From offshoring production over a decade ago to reducing conference attendance and author advances to cutting copyediting, fact checking, advertising, and worst of all, staff, publishers have spent these last two decades trying very hard to increase their margins. It wasn’t always this way.
Publishing used to be pretty swanky back in the last century, but then something changed: Amazon entered the building.
Books were always a means to an end for Amazon, a loss leader that opened the door to transforming retail across the globe. Readers have gotten used to getting their books fast and cheap, and that puts a tremendous amount of pressure on book publishers.
So it’s no surprise that publishers are turning to AI for book creation and backend streamlining. They will use it to estimate print runs, do financial forecasting, and create book P&Ls, and they’re already hiring AI engineers to help them implement these plans. Jacket design and social media marketing are an easy first step. Neither requires putting the book’s full content into a GPT (although we know publishers are doing this—more on that in a minute). For academic publishers who were already using a handful of templates for their monograph publishing programs, it would be a simple way to completely dispense with human jacket and interior designers on those lists.
Meta has publicly committed to policies asking accounts to indicate when content has been AI created or edited, but I don’t see publishers checking that box when they post. They are using AI to help them create their content, but they aren’t declaring it.
Our team attended the Publishers Publicity Association meeting earlier this year where AI was the topic of the day. I wasn’t too surprised by the ways they were using AI, but I was surprised by the extent. It sounds like, basically, for everything. They shared that they are uploading manuscripts into ChatGPT, which was very surprising. They are creating individual projects for each book, uploading the manuscript, and having it create 60-second pitch scripts for media meetings, first drafts of press materials, author Q&As, identifying target audiences, finding specific data within the manuscript like dates or character names, and lots more. If my brain could have gone “boom,” it would have. I was deeply shocked.
Here’s the part I keep coming back to: those are authors’ manuscripts. Some of them are books that haven’t been published yet. These are enterprise licenses and I am sure they have settings that prevent them from being used for training but how quickly we just gave up when lawsuits are still ongoing. We just talked about books being stolen to build these tools, and now some publishers are handing them over voluntarily. I don’t know whether the authors of those books were asked. I’d guess most of them weren’t.
Reading the book is the most critical task we do as publicists and marketers. If in-house teams no longer have time to do it, AI will certainly help them keep up, but that says more about publishing and continued cuts to workforce than it does about AI.
Fewer people working on more books, with AI filling the gaps, and real questions about what that does to the work.
What I thought was most interesting about the PPA session was the admission that using ChatGPT was dulling their skills, and that when ChatGPT is down, the publicists who were speaking had a harder time doing their jobs. I don’t know how long they’ve been using it, but this doesn’t bode well for our in-house colleagues and the skills training of junior team members and the next generation of book publicists. I worry that we may not need those junior team members at all if one publicist can upload a manuscript, have a GPT do all of the above, and just press go on a campaign. Could one publicist handle a hundred books like that? Probably, maybe more. Can they handle it well? No, I think not.
Most publishers have always leaned too hard on publicity as the main driver of book sales, because really good marketing is time consuming and can be expensive. But social media has destroyed attention spans, and those big book reviews and interviews with authors don’t carry book sales as far as they used to. Without significant changes in how marketing is conducted in conjunction with excellent publicity, we’re going to see individual sales continue to diminish.
With continued reductions in workforce, publishers can use GPTs to help marketers and publicists save time identifying audiences of readers based on a book’s subject matter, style, and genre. AI could be really helpful in refining metadata and making product search more effective. Using the same dozen rinse-and-repeat book marketing tricks isn’t working anymore, and meeting readers where they are is critical. This is work we have time to do as independent publicists and marketers, but our in-house colleagues don’t have that kind of time and creative space at their disposal.
We have already started to see significant AI uptake in publisher marketing in 2026. I’ll be curious to see how that works for them in execution and book sales. You can be sure I’ll be watching closely and talking to our in-house colleagues about it regularly.
Ask before your manuscript goes anywhere: will it be uploaded to an AI tool, which one, and can it be used for training?
A few questions worth putting to your editor or agent:
None of these are rude questions. You’re allowed to ask what happens to your own work. And it’s one more reason independent publicity matters more now, not less. When we take on a book at Page One, we read it. All of it. That’s where the best creativity come from. We’re not experts in your subject area, we’re generalists here and have worked on hundreds of topics. How do we learn how to pitch your book or build the right media contacts without reading? And, more importantly, we’re all here because we love to read, we love to learn, and we are the helpers. Delivering books into the hands of the right media helps us reach the right readers. To do that, we need to dive head first into the work.
Sarah Russo is the founder of Page One Media® and Page One Education™, the host of The Publishing Playbook™ podcast, and a publishing professional for over twenty-five years working across multiple facets of the industry. You can connect with her on BlueSky @sarahrusso.bsky.social, Instagram, Substack, and LinkedIn. You can follow the work of Page One Media on LinkedIn, @pageonem.bsky.social on BlueSky, Substack, Instagram, YouTube, and TikTok.