How AI Is Polluting the Publishing Industry

Artificial intelligence has made the dream of publishing a book easier to achieve than ever. It has also led to a surge in what some call “digital slop,” creating new hurdles for publishers, editors, and authors.

Insiders warn that the influx of AI-authored content is leading to market saturation and declining visibility for human authors.

Last year, more than half of novelists surveyed by the University of Cambridge in the UK expressed fear that their work would be replaced entirely by AI. And one publisher based in Seoul, South Korea, released about 9,000 books in one year, raising suspicions that most were AI-created to some degree, according to a February report in The Korea Times.

In the United States—the world’s largest book market—more than 4 million titles were published last year. From that pool, the exact number that were either generated or assisted by AI tools is unknown. A 2025 BookBub survey of 1,200 authors revealed that about 45 percent said they were using AI for writing, marketing, or other aspects of their artistic process.

“The economics of publishing are definitely shifting,” Nick Blewitt, an author and marketing consultant at The Book Marketer, told The Epoch Times. “AI has lowered the barrier to producing books quickly and cheaply, leading to a huge increase in content on platforms like Amazon. For authors, this means more competition and greater difficulty standing out.”

Blewitt said the challenge is no longer simply publishing a book, but proving originality, expertise, and authenticity.

This is particularly problematic on self-publishing platforms such as Amazon Kindle Direct Publishing, where authors complain about low barriers to entry and algorithms that reward speed and volume over quality.

An Amazon spokesperson told The Epoch Times: “We have content guidelines governing which books can be listed for sale, and have proactive and reactive measures for preventing, detecting, and removing content that violates those guidelines, whether AI-generated or not. We continue to enhance our protections against noncompliant content, and our process and guidelines will keep evolving as the technology and industry evolve.”

Copyright Issues

Authors’ use of AI tools isn’t the only problem. Publisher and editor use of the technology is also controversial due to potential copyright issues. In August 2025, tech giant Anthropic agreed to pay out the largest copyright lawsuit settlement in U.S. history—$1.5 billion—over the illegal download of more than 7 million books to train its popular AI model Claude.

“There are important legal and copyright issues that remain unresolved, including training data, ownership of AI-generated content, disclosure requirements, and liability for inaccurate or plagiarized material,” Blewitt said.

A 2026 BookNet Canada survey of the North American book industry found that 46 percent of respondents admitted to using AI at an individual level, and 48 percent said their organization uses AI tools. Eighty-six percent identified inadequate controls surrounding copyrighted material as their main concern about AI in the book industry.

The real issue for publishers and authors is not whether but how AI was used, said Christina Johnson of Testimony Publishers.

“One major misconception about AI-written content is that the person who ‘feeds’ AI the data automatically owns the output,” she said. “That assumption is most often wrong because there’s a difference between AI-assisted and AI-generated.”

It’s a legal distinction now at the forefront of copyright discussions. U.S. copyright law protects only works that are “created by a human being,” according to the U.S. Copyright Office. However, the degree of human authorship required is still a gray area. The issue is nuanced and is being considered on a case-by-case basis.

Currently, someone who uses an AI platform to generate a book without being involved in the creative process has no copyright protection. However, if the human author makes some changes and adds some of his or her own creativity, the author may qualify for U.S. copyright protection.

Marketing

One might argue that poorly written AI book content will sink to the bottom of the sea of literature without creating any trouble for traditional authors.

Some writers say that’s not the case.

“The cost of marketing a book is 10 [times] what it was a decade ago,” international bestselling author and innovation expert Jay Samit told The Epoch Times.

Samit has worked with major publishing houses such as Simon & Schuster and Macmillan Publishers. He said that when it comes to marketing, the high volume of AI books has changed everything.

Years ago, as an author with an established following, Samit allocated 85 percent of his time to writing and 15 percent to marketing.

“In 2026, to maintain the same level of sales, the ratio has been inverted, with 85 percent of my efforts going toward marketing,” he said. “My prediction is that half of the independent publishers will go out of business with these economics and with consumers having to sort all of the AI clutter.”

Samit said he believes that the surge in AI books could actually translate to a decrease in overall sales.

“Publishers do virtually nothing to market a book, unless you are a major name. The majority of authors are on their own to get their books to sell through, which is why most of the millions of titles published each year sell under 100 copies,” he said.

Samit offered a glimpse into how much work he is putting into marketing his latest book.

“I have worked 50 [plus] hours a week appearing on dozens of podcasts, sending out query letters and press releases, building a website, training a custom AI that I’m giving away with the book, writing my Substack, giving virtual and in-person speeches around the country, making radio and television appearances, making video trailers, and posting daily across all social media channels,” he said.

“If I didn’t believe that my book was going to help people at a time when AI is going to cause millions to lose their jobs, I wouldn’t be doing this.

“I am lucky that I am not trying to make a living as an author; that business model is now broken.”

The Specter of AI Can Ruin You

The “Shy Girl” scandal illustrates how much AI—and the fear of its use in literature—can create complex issues.

Originally a 2025 self-published horror novel by Mia Ballard, “Shy Girl” quickly became an online sensation. Publisher Hachette Book Group picked up the novel, but it canceled the release earlier this year over suspected AI use in the manuscript.

Ballard reportedly denied using AI but said an associate hired to edit the self-published version of her manuscript may have used AI tools.

Ironically, the main source for detecting AI in writing is other types of AI, and it isn’t infallible. In one instance, Audrey Henson of The Drey Dossier tested an AI detection tool with a poem that she created years before large language models existed. Although the overall scan of the poem came back 91 percent human, when it examined individual lines, the AI tool claimed that the writing sample was 70 percent likely to have been written by AI.

The takeaway: Even if authors aren’t using AI in their writing, one mistake made by a detection tool could have major consequences for their credibility.

Readers Pushing Back

“AI is dramatically increasing the volume of written content. That lowers the barrier to entry but also flattens the average quality,” Belinda Hooke, founder of Acorn Studio Press, told The Epoch Times. “The result isn’t just more books, it’s more noise.”

Hooke’s publishing company focuses on “mission-based, paper-first” literature designed to help children develop critical thinking skills. She said she is seeing readers push back against AI book content and seek out trusted voices and community-driven recommendations.

“We’re seeing authors rise not because they can produce more, but because readers choose to carry them,” she said.

“One of the biggest misconceptions is that writing a book is the same as creating a meaningful reading or learning experience. Generating text is only a small part of that. Structure, pacing, and the way a reader’s thinking unfolds are much harder to replicate.”

Johnson shares that perspective.

“AI may reduce production costs, but it hasn’t decreased editorial review time, quality control issues, rights management, author instruction, and the list goes on,” she told The Epoch Times.

“AI can make manuscripts and books faster, but it can’t generate lived experiences, spiritual conviction, creative insight, or reader trust.”

Johnson said she has seen a sharp uptick in AI-assisted and -generated manuscript submissions over the past 18 months.

“Our team is seeing more and more authors using AI in their planning, drafting, and editing processes,” she said. “I’ve seen some of them use it responsibly, for brainstorming, research organization, and idea development. But I’ve seen more cases of authors using it in ways that raise my concerns.”

Reshaping Accountability

“AI will reshape the volume of publishing, but more importantly, it reshapes accountability,” Hooke said. “When anyone can publish anything, the role of a publisher shifts from gatekeeper to guarantor of quality, trust, and relevance.”

In March, The Author’s Guild launched a “human authored” certification program for writers and publishers who hope to “distinguish their human-written books from the glut of AI-generated books already saturating the marketplace.”

Blewitt said he remains hopeful that the industry will adapt, much as it has with e-books and self-publishing over the years.

“Human-led storytelling and authentic expertise may become even more valuable as readers grow more cautious about mass-produced content,” he said.

Johnson said she thinks that the focus should be on accountability instead of an all-or-nothing approach.

“As publishers, we don’t need to treat AI as ‘forbidden fruit,’ but we will need to use editorial discernment, require author transparency, and a clear standard for protecting our voices,” she said.