AI adoption in publishing is widespread, which you wouldn't expect given the tenor of the AI conversation among publishing professionals. The uptake is happening. But it's an uneasy, undocumented uptake—and no one is comfortable with the situation.

According to BISG's 2025 survey of more than 500 North American publishing professionals, just under half of individuals and organizations are actively using AI tools for work. Nearly sixty percent said they stay informed about AI, even if they haven't committed to using it yet. And nearly 100% have concerns. These numbers tell a story: publishing leadership is paralyzed and falling down on the job of creating an AI policy in publishing.

The Book Industry Is Using AI, Whether It's Ready or Not

One thing publishing personnel can agree on is that the way AI uptake is going is giving them heart palpitations: ninety-eight percent of respondents to BISG’s survey reported having ethical concerns or pain points about AI. Nearly all of them. Adoption is climbing (show yourselves!). But trust that we can wield this tool in a way that makes us proud is on the floor.

That contradiction is the defining feature of this moment. AI adoption in publishing has arrived faster than the industry's confidence in its own judgment about how to use it. Most of the discourse has been focused on suspicion of authors and AI-written books. But what about publishers (who do, after all, have the lawyers) telling authors how they’re using, thinking about, and planning to use AI? AI is here; its use cases are multiplying. But where is the leadership towards how AI should or should not be used?   

AI can shave hours off metadata generation, back-cover copy, and first-pass editorial notes. It can also introduce errors, legal exposure, and an erosion of the author relationships that publishing depends on. There are pros and cons, in other words. Pretending we can let this wave crash over our industry and our creators and sort it out “later” is just paralysis. And we will regret that.

What Does the Adoption-Comfort Gap Actually Look Like?

The BISG data breaks the industry into clear segments. Service providers are adopting AI at significantly higher rates than publishers or libraries, with individual adoption at 82% compared to roughly 43% for publishers. They also have clearer internal governance: zero percent of service providers said they didn't know whether their organization had an AI policy. So it looks like it’s not impossible to create these policies, Publishing.

Publishers, by contrast, are split almost evenly across three buckets:

• Organizations that have a policy in place and can point to it.
• Organizations that don't have any stated position at all.
• Organizations currently developing one, often reactively, after a specific incident forced the question. Don’t let this be you!

Libraries show even less internal structure. The gap between who's using AI and those who have guidance around it is the structural risk hiding in plain sight, and it explains why 98% harbor serious concerns despite, clearly, some of them using AI. The most dangerous thing for everyone involved is that the tools are in daily use, but no one knows the rules.

Why the Most Common AI Concerns Are Publishing-Specific

The top concerns in the BISG survey aren't abstract. Ranked by how many respondents flagged them, they are:

1. Inadequate controls around copyrighted material: 86%. For an industry whose entire economy is built on licensed rights, this is closer to a business-model threat than a technical inconvenience.
2. Hallucinations: 84%. A fabricated fact in a nonfiction title or a misattributed quote in marketing copy is a reputational liability that outlives the sentence that caused it.
3. AI-generated books flooding platforms: 81%. Discoverability is already a battle; a tide of low-quality synthetic titles makes every real author's book harder to find.

These aren't hypothetical worries; they're existential ones for an industry built on intellectual property, editorial authority, and reader trust. 

What Publishing Leaders Need to Decide Right Now

The BISG survey asked respondents where guidance was most needed. The priorities landed in a clear order:

• Metadata: 48%. The area where AI is both most useful and most likely to introduce silent errors at scale.
• Laws and regulations: 40%. A fast-moving landscape where standing still is itself a risk.
• Editorial guidance: 39%. The heart of the craft, and the place authors care about most.

A clear position tells your team what AI is for in your house, what it's not for, and who's accountable for outcomes. Even a stance of "we are not using AI at this time for ethical reasons" is legitimate as long as it's stated, documented, communicated, and enforced. The organizations that navigate this period best are the ones that get ahead of the ambiguity before it creates inconsistency, liability, or loss of trust with authors and readers.

In practice, we recommend starting narrow and honest. Pick the one workflow where AI already touches your work, usually metadata or marketing copy (good news; BISG found these are the places teams are interested in using AI!); write down what's permitted and what isn't, name the person who reviews the output, and tell your authors what you decided. A one-page policy that people actually read beats a twenty-page framework that lives unopened in a shared drive. A written AI position doesn't slow a publishing house down; it's what lets the house move without stepping on a landmine.

The data tells a coherent story: publishing wants the efficiency AI offers and doesn't yet trust itself to use it well. 

At Next Chapter AI, we believe human craft stays central to every good book, and that AI's proper role is to remove the drudgery that keeps editors from editing and authors from writing, never to replace the judgment that makes a book worth reading. The houses that thrive in this next chapter will be the ones that pair enthusiasm for what these tools can do with the honesty to say plainly what they won't.

If you're ready to build that kind of clarity, a position your team believes in and your authors can trust, we'd be glad to help you write the first page of it.