Draft2Digital rejects between 40 percent and 75 percent of all book submissions—and it’s almost entirely AI-generated nonfiction. Meanwhile, fiction authors using AI tools as creative assistants are doing just fine. That split gives authors important context in the ongoing debate about whether AI will disrupt the independent publishing industry.
Draft2Digital CEO Kris Austin has a thirty-thousand-foot view of this landscape. His company distributes books from roughly 330,000 authors—approaching 1.2 million titles—to retailers worldwide. He’s watched the AI flood firsthand, and his assessment cuts against both the doomsayers and the utopians: People consistently overestimate what large language models, such as ChatGPT, can do.
“I would just recommend people take a bit more of a middle-ground approach to how they think about these tools,” Austin says. “This technology is really cool. It does really neat stuff. But it has limitations.”
Those limitations matter more than most people realize, and they cut in directions that should reassure fiction authors—but should warrant caution in situations where information quality matters.
The Real Crisis: Nonfiction Flooding the Market
The nonfiction side of publishing is fighting an active war against AI-generated content, and the scale is staggering. Thousands of AI-generated books arrive at Draft2Digital every single day, Austin says, and the company turns most of them away.
This problem predates AI. A decade ago, Draft2Digital was fighting off literally thousands of smoothie recipe books uploaded by people chasing quick money, Austin says. Whenever something gets popular—keto diets, spirituality, cryptocurrency—the opportunists swarm in with low-quality content designed to ride the trend. The difference now is speed: What used to require at least some human effort can now be produced in minutes.
“This is not a new problem,” Austin says. “However, AI tool sets can create these sorts of nonfiction books in mere minutes.”
Some of this content poses genuine danger to readers. Austin points to the mushroom-foraging books that made news recently, in which AI-generated guides confidently declared certain mushrooms safe to eat when they were actually poisonous. “There’s a lot of dangerous content that can be generated by these LLMs,” Austin says. “You’re going to continue to see retailers push back on that sort of content one way or another, because it’s going to risk liability.”
Draft2Digital watches for certain telltale signs in the accounts it flags, including when an account publishes across dozens of unrelated nonfiction topics. When someone puts out books on nutrition, spirituality, business strategy, gardening, and cryptocurrency all at once, red flags start waving.
“We’re kind of like, ‘Do you know anything about any of these topics?’” Austin says. The answer is almost always no. Publishers used to pay ghostwriters to produce content they knew nothing about; now, some have cut out the ghostwriter and let the LLM handle it directly.
What Retailers Are Actually Worried About
When Amazon introduced its checkbox asking whether content was AI-assisted, speculation ran wild about what it meant for fiction authors. Austin offers a clarifying perspective: Retailers aren’t primarily worried about fiction authors using AI tools to brainstorm Romances or clean up Thrillers. Their concern is the nonfiction flood.
“They’re more interested in filtering out the nonfiction work because it’s just completely flooding their interfaces, their websites,” Austin explains. “It’s the nonfiction that is a real challenge for the industry.”
This reframing matters for indie fiction authors who’ve been anxious about AI disclosure requirements. The target isn’t the Romance author who uses ChatGPT to brainstorm character backstories—it’s the account churning out fifty dubious self-help books a month.
Austin expects more verification processes around expertise and credentials for nonfiction in the coming years. If you’re publishing a book about medical advice, financial planning, or which mushrooms are safe to eat, retailers and distributors increasingly want assurance that you actually know what you’re talking about.
On a more positive note, Austin sees retailers potentially using AI to improve discoverability—helping readers find books they’re actually looking for amid the ocean of available content. Traditional algorithms might give way to more sophisticated AI-powered recommendation systems, meaning the same technology creating the problem could help solve the challenge of connecting readers with quality books.
Fiction Authors Can Relax (Mostly)
Austin doesn’t believe AI-generated novels will steal readers; something essential disappears when a machine writes a story, he says.
“Some authors have attempted to write full stories using AI tools. They tend not to do very well in the market,” Austin observes. “That humanity seems to be missing. It’s kind of one of those things you can’t directly define, but it doesn’t appear that readers are connecting with that content.”
Most indie fiction authors are using AI the way they’ve used other tools for decades—as assistance rather than replacement. They brainstorm plot points, research historical details, clean up grammar, and smooth rough edges. AI has become another tool in the belt, sitting alongside other helpers authors have embraced over the years.
“I would just recommend people take a bit more of a middle-ground approach to how they think about these [AI] tools.”
—Kris Austin, Draft2Digital CEO
“Where they [some authors] may not be able to afford copy editors or line editors, I think these tools can kind of help them smooth out or recognize some of the rough edges they otherwise couldn’t have solved,” Austin says.
Austin predicts that fully AI-generated fiction will crash and recede as a trend on its own. Readers want connection, and they want to know a human being wrestled with those words, made choices about which scenes to include, and decided how the story should end. He believes readers will, at first, be “excited they [AI models] can create fiction. They might read one or two of them that way, and the next thing they know, they’re just buying human-authored books again,” Austin says. “Readers would feel betrayed if a book was sold as human-authored and turned out being AI-authored. They’re going to be pretty much furious.”
That fury will act as a natural market correction, he predicts. Readers themselves will drive what works and what doesn’t, and authors will fall along a spectrum—some avoiding AI entirely, others using it to the maximum—while the market sorts out which approaches resonate.
Draft2Digital’s Role
Draft2Digital has positioned itself as a gatekeeper for indie content, a role that generates friction but one that Austin sees as essential for the long-term health of indie publishing.
“We want to help retailers have the best indie books available, and we don’t want them to be worried that indie content is going to destroy their storefronts,” Austin says. “We think it’s part of our job to help keep the indie author reputations at a high level because there’s just so many good books.”
The work is unglamorous, and it makes people angry when they disagree with rejection decisions. Austin argues the industry doesn’t have much choice, though—if thousands of low-quality books flood the market and overwhelm retailer storefronts, the damage extends to every legitimate indie author trying to reach readers.
Beyond filtering, Draft2Digital actively works with retailers on how author content gets used, particularly around AI training. When retailers want to do something with content that makes Draft2Digital uncomfortable, they push back and try to find ground that accomplishes business goals without compromising author rights. “We want to protect authors’ rights when it comes to how their content is used, that it’s not trained by LLMs without their consent,” Austin says. “We’re working with the retailers to help find that balance.”
The Long View from Someone Who’s Seen It Before
Draft2Digital exists because technology opened doors that were previously closed. Ebooks made publishing accessible, and print-on-demand eliminated the need for warehouse investments. The company was founded in 2012 specifically to help authors navigate the technical complexity of digital publishing.
Austin’s journey to CEO wasn’t exactly planned, either. A friend who was an author needed help getting his books formatted for distribution, and a mutual friend wrote software to handle the technical parts. By 2012, Austin was being asked to run the resulting company. He talked to his wife; quit his job with four months of runway, a three-year-old son, and a six-month-old daughter; and hoped it would work out.
“If I had to do it over again, I wouldn’t do it,” Austin admits with a laugh. “Being naive about running a business back then is the reason why I was willing to do it. The risk that we took—I didn’t fully recognize how risky it was.”
That naivete paid off. The company grew rapidly, acquired Smashwords in 2022, and became the dominant player in wide distribution for indie titles. They’ve expanded into print-on-demand, making it easy for authors to convert ebooks into print books with automated tools, and more formats are coming in 2026.
Throughout it all, Austin has watched technology transform publishing while keeping perspective on what technology can and cannot do. He’s seen the smoothie book floods and the keto book floods and now the AI book floods, and he’s watched authors adapt to ebooks, to algorithms, to social media marketing, and now to AI tools.
One thing he keeps coming back to: Indie authors are remarkably good at adapting. They respond quickly to changing trends, find creative solutions, and tell stories that serve specific niches traditional publishing won’t address. That adaptability has always been indie publishing’s defining advantage, and Austin doesn’t expect AI to change that fundamental dynamic.
“I think it just helps indie authors be a bit more responsive and to kind of craft some better works from an end product,” Austin says of AI tools in fiction. Authors who use these tools thoughtfully, as assistance rather than replacement, will find them valuable, he says. Those who try to outsource their creativity entirely will discover that readers notice and care.
The Bottom Line
A peek behind the curtain at Draft2Digital shows that both the apocalyptic predictions and utopian promises about AI in publishing made at the inception of the technology were wrong. The technology is hitting limitations—the rapid improvements of the early years have slowed. Readers still want human connection, and the market still rewards authenticity. The nonfiction flood is a serious problem requiring serious responses, but it’s ultimately a problem of scale and speed rather than a fundamental transformation of the industry.
“We’ve all seen it. They’re not getting better and better as they were the first few years. They’re slowly getting better now,” Austin says.
For indie authors, Austin’s advice is straightforward: Keep doing what you’ve always done. Tell your stories and use the tools that help you tell them better. Trust that readers want what you’re offering not because AI can’t generate words—it obviously can—but because those words aren’t the same as yours.
The “middle-ground approach” he recommends isn’t the most exciting position to take, and it may not generate breathless headlines or viral social media threads. It might, however, be the most accurate assessment of where the industry actually stands and where it’s headed.
