Perhaps no topic in publishing today is more contentious than the use of artificial intelligence, especially generative AI. As the technology develops at lightning speed, questions of copyright, fair use, and long-term consequences have yet to be fully addressed. How are publishers to weigh the opportunities AI presents against the ethics involved in its use?

Opposition and Opportunity

Some publishers oppose nearly all use of generative AI. “I abhor the idea of using AI for anything other than as a tool to provide analyses of complex problems in simple terms,” says author-publisher Mark Richter. “Writing and publishing do not fall under these terms to me.”

Wiley Barnes, director of Chickasaw Press, says he sees the potential for using AI in marketing, but beyond that, his company steers clear of it. “AI is problematic in publishing and producing accurate and inclusive accounts of our Chickasaw history and culture,” he says. “We have authors who have put in years behind this effort, and [we] will not participate in the copyright and intellectual property being used illegally and in an inappropriate way.”

Electronic publishing analyst and author Thad McIlroy, president of The Future of Publishing and author of The AI Revolution in Book Publishing, notes that discussions of AI in publishing tend to be highly inflammatory and highly emotional. “Nuanced perspectives are not perceived in this debate, for or against,” he says.

He suggests that once the courts have definitively settled copyright questions involving AI, the topic may become less contentious. “In the meantime, I think we’ll see continuing animosity on both sides,” he says.

But definitive court rulings could be years in the making. Meanwhile, AI technology is developing at a rapid pace. For publishers, that means something of a balancing act between opposition and opportunity, with those who reject AI wholesale potentially putting themselves at a competitive disadvantage, McIlroy says.

At Ulysses Press, CEO Keith Riegert believes the benefits of AI for publishers “vastly outweigh” the negative impacts. “Models like ChatGPT, Claude, Gemini, and Copilot have transformed the level of efficiency and productivity we can achieve in publishing,” he says on his AI webpage. “And this is the least ‘intelligent’ AI we will use for the rest of our lives.”

Arete Media International’s CEO Aisha Jackson says that, for her, AI has become indispensable. “AI has completely transformed my business, turning it from a struggling operation into one with a solid foundation,” she says. Jackson uses AI to transcribe dictated thoughts, generate copy for sales funnels and internal communications, and map out processes in her back-end operations. “Operationally, AI has been a game-changer, not only relieving me of many time-consuming tasks but also allowing me to implement processes and systems that I simply wouldn’t have been able to manage on my own,” she says.

With AI, Jackson says she can readily repurpose material from books, workshops, and videos into new formats such as courses and e-books. She also relies on AI for editorial tasks, prompting it to suggest relevant quotes and current examples that she says keep her content fresh and engaging.

All in all, Jackson says her use of AI has resulted in a stronger and more sustainable income stream. “The success I’ve seen in efficiency, client satisfaction, and overall growth speaks for itself,” she says. “Embracing AI has been one of the best decisions I’ve made.” In a previous career, author-publisher Robert Muller of Poesys Associates built machine-learning systems. Now, he uses Perplexity, ProWritingAid, DeepL Write, and, to a lesser extent, ChatGPT, for brainstorming story ideas, content editing and critiques, line editing, copywriting of blurbs, and generating book descriptions, loglines, and marketing copy.

“I can’t tell you how much time and effort I’ve saved,” he says. “AI takes the pain out of writing novels and lets me focus on the writing rather than on the myriad details of book production and marketing.”

Efficiency, but at What Cost?

AI tools abound, offering efficiencies with the potential to benefit a publisher’s bottom line. Riegert cites a recent Harvard Business School report showing that with AI assistance, consultants worked 25% faster with 40% better results. “If this holds true, this represents a revolutionary jump in efficiency and productivity in the history of work,” he says.

On his AI webpage (perfectbound.co/page/ai), Riegert suggests 20 ways publishers can experience these efficiencies for themselves. Need a marketing plan for a new release? Ask ChatGPT to create one. How about an image for social media testing? Midjourney can help. Want to clean up an author photo? Try Photoshop beta. A book trailer from text? Wave.video can make it happen. Looking for a summary of a manuscript? Anthropic’s Claude can generate one in minutes. Tracking your digital marketing efforts? Google Worksheet Duet can set up a system.

In researching his book, McIlroy identified far more AI tools focused on writing tasks than on publishing outcomes. Of the publisher-specific tools, he says Veristage’s Insight is “the most far-reaching and ambitious” for both editorial and marketing tasks, while Shimmr harnesses the power of AI to create ads and direct them toward the right audiences.

Jackson relies on Claude, ChatGPT, and Google Gemini for her creative and generative tasks. She also uses ChatGPT on the back end to map out processes. “This streamlines my operations and ensures that nothing falls through the cracks,” she says. To analyze customer feedback and categorize content, she uses MonkeyLearn.

In the author services space, PublishDrive’s community manager Danica Favorite says the company uses an AI chatbot on its website to answer customer questions. “We’ve also come out with an AI product that helps authors and publishers write book blurbs and choose the correct Amazon categories and metadata for their books,” she says. “Overall, we consider our efforts to be quite successful, and we plan on doing more in the future to incorporate AI into our business.”

At Launch My Book, President Joel Pitney uses Shimmr to create and manage digital ads. “Using AI for our ads has dramatically lowered our costs so that we can deliver ad services to authors with a much better chance of [them] paying for themselves in terms of royalties,” he says. His company also uses Opus Clips for video editing of social media clips and ChatGPT for drafting marketing copy.

After years as an independent publisher, Gavin Marcus joined forces with author and behavioral finance expert Jeremy Esekow to create Storywise, an AI-powered tool for authors and publishers. “Our platform helps publishers quickly discover authors who have written manuscripts that fit their wish lists,” Marcus says. “We prioritize the most promising content, freeing up huge chunks of the working day for value-added work.”

But publishers also worry that as they relegate tasks to AI, jobs will be lost. “It's going to be a quiet bloodbath when it comes to human staff,” Riegert says. “It’s already started.” Even so, McIlroy suggests that fewer jobs will be lost to AI in publishing than in other industries. “For every job lost, another is created,” he says.

That doesn’t diminish the fact that employees are scared, says Nabil El-Ghoroury, an ethics expert who worked on publishing projects for the American Psychological Association (APA). He points to data from the APA’s Work in America survey suggesting that employees of color worry more about AI replacing their jobs than others do. Workforce development is essential to transition staff into new positions that will arise as publishers adopt AI, he says.

What of authors who fear that AI will make their creative work irrelevant? “A publisher could be incentivized to accept machine-generated content,” McIlroy admits. But he points out that if authors look at the AI-generated competition on Amazon, they’ll see it’s not getting on the charts. “AI could replace authors conceptually,” he says. “But in reality, it’s nothing like that.”

Marcus notes that content generated wholly by AI can’t be copyrighted in many countries. “AI cannot replace a good author,” he says. “It has no soul or lived experiences.”

At the same time, he says publishers should encourage authors to use AI tools to assist in their creative work. “Better to be ahead of the curve than to fall behind and to become uncompetitive,” he says.

Ethics Matter

The use of AI in publishing is fraught with ethical dilemmas. “When I first considered using AI in my business, I was concerned about whether it felt like cheating or taking a shortcut,” Jackson says. “I’ve always believed that shortcuts lead to long roads, and I was wary of compromising the authenticity and quality of my work.”

In the end, she decided she could not afford to ignore tools and technology that could make or break her business. When working with clients on manuscript development, she makes sure her clients know how she uses AI. “Transparency has been key, and it’s something I’ve made a cornerstone of how I operate,” she says.

Pitney agrees with the need for transparency. “We tell everyone upfront about where we use AI and for what, and [we] get their buy-in before proceeding with any work contracts,” he says. “I think ethical issues arise when content and services powered by AI are being sold as otherwise.”

Compensation for creative work is another basic ethical consideration surrounding AI use, McIlroy says. “The major ethics are whether or not the AI companies have the right to accrue value from a work without compensating the individuals,” he says. “Fair use doctrine in law may insulate the AI companies from having to face the ethical dilemma.”

Or not. Marcus notes that litigation over potential copyright breaches by Stability, Anthropic, and OpenAI is ongoing. “There have been no adverse rulings yet,” he says. “And when there are rulings against these frontier models, compensation arrangements will come. This is what happened in May 2020 with Spotify and music creators, resulting in a massive boost in music creativity and discovery.”

When the courts rule, Marcus expects authors and publishers will have the option to withhold their material from training AI models. But he warns that those who do so will risk being left behind. “The models are already too advanced, and AI adoption in general is well past the tipping point,” he says.

When weighing ethical concerns, El-Ghoroury suggests publishers identify the circumstances under which they intend to use AI and remember the importance of keeping a human in the loop. Publishers should also consider the ethics of feeding content into nonproprietary platforms such as ChatGPT, he says. “If you’re using AI that’s commercial, you’re putting someone else’s intellectual property into a system,” he says.

Even as the AI landscape evolves, publishers can develop policies around using it ethically. “People are always afraid of how to write a policy and cover everything,” El-Ghoroury says. “Start with something and update it.” By gathering groups of peers to approach the issue from different angles, associations like IBPA can help, he adds. For fine-tuning policies, he suggests a critical incident analysis of problematic examples to identify gaps and key questions.

For help in navigating questions of AI and ethics, McIlroy suggests George Washington University’s Ethics in Publishing conference. And as Pitney points out, even casual discussions among industry professionals can be useful. Toward that end, Favorite moderates an online Facebook group called AI for Authors. She also co-hosts a podcast, Brave New Bookshelf: Perspectives in AI and Publishing.

“It’s important to keep learning and following what's happening in the industry,” Favorite says. “Things will continue to develop faster, and already AI is doing more than we anticipated, so it’s important to keep following the developments.”

McIlroy urges publishers to stay aware of the fast-changing landscape and not box themselves into a corner. Rather, they should create policies that are far-reaching on principles, such as supporting creativity and authors. Wholesale rejection of AI in creative processing may sound good at the time, McIlroy says, but because we don’t yet know all the ways in which AI can benefit authors, painting policies with a broad brush may end up working to their detriment.

A Changing Landscape

Amid the uncertainty that AI brings, one thing is certain: the changes and challenges will continue.

“If you think AI today is on par with a first-printing-press-level change, the next generation of AI is going to be like going from Gutenberg to the MacBook in a matter of months,” Riegert says on his AI webpage. “This environment is going to keep shifting at speeds we humans are not really equipped to handle. The best thing you can do is keep a close eye on what’s next, explore constantly, and use AI every single day.”

Jackson takes this advice to heart. “As I look to the future of AI in publishing, I know that staying adaptable is going to be key. The way AI is evolving, along with the ethical questions that come with it, means I’ll need to keep learning and adjusting as things change. My goal is to use AI to enhance what I do without losing sight of the principles that matter to me and my audience.”

On a broader scale, Marcus urges publishers to go beyond trying to cope with these changes and join the discussion on the evolution—and revolution—that AI brings. “We are actually the gatekeepers,” he says. “We often decide what gets published and ultimately what the [large language models] currently reference and what they will train on in the future.”

Weighing the opportunities AI affords against the ethical concerns it raises is no simple task. But by staying informed, engaging in meaningful discussions, and developing policies that can evolve with the landscape, publishers can position themselves to thrive

---