Why Book Licensing Is Harder Than Anyone Thought

Most people think licensing a book for AI use is simple: find the book, find the author, pay the author, done. That theory breaks the moment you actually try to do it.

Trip Adler has tried. He’s now built the infrastructure to handle what the entire industry decided was impossible — clearing rights on millions of books where each book has between three and seven different people who need to approve and get paid.

A bestselling novel isn’t owned by one person. It’s owned by a network of stakeholders, each with different rights, different incentives, and different power. Ignore one of them and your licensing deal doesn’t work. The AI company can’t legally train on the data. The book can’t be used. The whole marketplace collapses before it starts.

This is the problem nobody solved before Trip tried. It’s also why most AI companies decided to train on whatever data they could find first and negotiate licensing later (or never).

How Many People Actually Own a Book?

Here’s what you need to know about how book rights work: one book isn’t one rights holder.

“When it comes to trade books — the books you see at the bookstore — the rights get pretty complicated and there’s multiple rights holders,” Trip explains. “For a particular book, you typically need to get sign-off from the author, maybe the co-author, the publisher, sometimes there’s multiple publishers involved, there’s often a literary agent involved.”

Let’s say you want to license a bestselling novel for AI training. You need:

1. The author (or authors if it’s collaborative)
2. Any co-authors who have stake in the work
3. The primary publisher
4. Sometimes a co-publisher (different territories have different rights)
5. The literary agent (who often controls subsidiary rights like foreign editions)
6. Any estate holders if the author is deceased

For each of these parties, you’re negotiating:

• Whether they even want to license to AI
• Which rights they’re comfortable licensing (training vs. reference vs. transformative)
• How much revenue they get from the deal
• How the revenue from one license gets split across multiple parties

This isn’t a licensing problem. It’s a legal, operational, and financial coordination problem at scale.

Why This Broke the Industry

When AI companies started building, they looked at this complexity and did the math. To train on a million books, they’d need to clear rights on a million books. That’s potentially millions of separate negotiations, each with different stakeholders.

“They looked at how complicated the rights were and they’re like, this would be impossible to clear the rights for, because they all needed a lot of books, like a million books, it was just impossible to clear the rights for the books,” Trip says.

So they didn’t try. They trained on whatever data was available online. They relied on fair use arguments to defend the practice. They waited for the lawsuits.

This wasn’t immoral. It was rational. The transaction costs of clearing rights on a million books legitimately was higher than the expected cost of litigation. Until it wasn’t.

What Created by Humans Actually Solved

The genius move wasn’t to make licensing easier. It was to make licensing scale without needing to contact each party individually.

Here’s how: built a platform where individual authors can claim their works, verify their identity and ownership, select which rights they want to license, and then execute a licensing agreement with one click.

The platform handles the multi-party complexity on the backend. When an AI company licenses a book, the system automatically distributes payment to every stakeholder according to their rights agreement.

But that only works if you can verify ownership. “We built a way to kind of scalably get these rights,” Trip explains. “An author can just sign up, verify their identity, select the rights they want to license, and then with one click, sign a licensing agreement.”

This is deceptively simple. The work is in the verification, the payment splitting, and handling the edge cases where publishers still own rights even when an author wants to license.

The Competitive Moat

This multi-rights solution is why Created by Humans has a defensible position even as bigger competitors (Microsoft, Amazon, others) enter the licensing space.

Most of these competitors approach the problem from the publisher side. They negotiate with major publishers, clear the rights at that level, and get access to broad catalogs. That works for publisher-owned rights.

But trade books — the books consumers actually buy and read — have fragmented ownership. A publisher controls the US rights. An agent controls the UK and foreign rights. Authors control subsidiary and direct-to-consumer rights in some cases. No single publisher controls everything.

“We’re the only company to take that particular approach to the market. So I think for that particular niche, we’re sort of, you know, we’re really kind of in a leading role for that,” Trip says. “We’ve kind of gone the furthest in solving this particular problem.”

The fact that this was “impossible” six months ago and now has a working solution matters. It means the barrier to entry just got lower. It means licensing suddenly became technically feasible at scale.

What This Means for AI Companies

For founders building AI products, this changes the timeline. Two years ago, licensing books seemed legally gray and operationally impossible. Now it’s operationally straightforward (thanks to Created by Humans) and the legal greyness is being clarified by ongoing lawsuits.

Trip’s bet is that once courts rule on fair use — especially if they rule that fair use doesn’t cover AI training on books — companies that already have licensing relationships won’t have to scramble to retrofit them. They’ll already be inside the system.

“We want to work with companies of all sizes and if you’re a small team and you want to try to do this, we’re very happy to work on it with you and find a model that works,” he says.

The timing matters because the companies that move first — that establish licenses before they’re legally required — will have faster scaling paths when licensing becomes mandatory. They won’t be negotiating with angry authors in a hurry. They’ll be expanding existing relationships.

The Inevitable Consolidation

What’s interesting about this problem is that it can only be solved once. Once someone builds the platform to handle multi-rights licensing at scale, that platform becomes table stakes.

Microsoft, Amazon, Google — they’ll notice this is now possible. Some of them will build their own. Some might acquire the team. But the outcome is inevitable: licensing becomes the default path for responsible AI companies.

Trip isn’t worried about competition because the market is so large that multiple players can win. “I mean, there’s so many companies, there’s so much going on in the world that there’s plenty of room for lots of companies here,” he says.

What he’s focused on is execution. Being the first to prove that multi-rights licensing works at scale, so that when the market inevitably moves this direction, Created by Humans is already the reference implementation.

FAQ

Why do books have so many rights holders when music doesn’t?

Music labels consolidate rights. A single label controls hundreds of songs and can license to Spotify. Publishing is distributed. Authors own their work. Publishers control editions. Agents control foreign rights. There’s no single consolidation point, so licensing requires contacting multiple parties per book.

Can an AI company just license from the publisher and avoid contacting authors?

Partially, but publishers don’t always control all rights. They control the edition they published. Authors often control digital rights, subsidiary rights, and direct-to-consumer sales. Different authors have different contract terms with their publishers, so the rights landscape varies per book.

What happens if an author wants to license their book but the publisher disagrees?

That depends on the publishing contract. Some authors have rights reversion clauses if the book goes out of print. Others have signed away digital rights to their publishers. This is why verification is so hard — every book has different contract terms.

How much does it cost to verify ownership of a book?

Created by Humans doesn’t charge authors to claim or list their books. Verification happens through author verification (matching identity to known databases), ISBN cross-reference, and publisher confirmation. The costs are built into the platform.

If I’m a co-author, how do I get paid for a licensing deal?

Created by Humans handles multi-party payments. Co-authors can both claim the work, set their ownership percentages, and the platform automatically splits revenue according to their agreement when the book is licensed.

What if a book has an agent who controls certain rights?

The agent can claim rights on their author’s behalf, or the author can claim the work and the system asks for agent consent based on the contract terms. It gets complicated, but that’s what the platform automates.

Can you license parts of a book instead of the whole book?

Potentially, but the agreements Trip has done so far are for full books. The system could theoretically support chapter-level licensing (for reference or transformative uses), but that’s not the current model.

Why is transformative rights licensing harder than training or reference rights?

Transformative rights involve AI modifying or creating derivatives based on the original work. Authors are more concerned about AI changing their work than AI referencing it. So the negotiation and verification layer is thicker — more authors say no to transformative rights than to other categories.

Does licensing cost the same for a bestseller as a niche book?

No. High-demand books command higher licensing fees because they’re more valuable for training. Niche books license cheaper. The market is discovering pricing as deals get done, but there’s no fixed rate yet.

What if a book is out of print — can you still license it for AI?

Yes, and actually more easily in some cases. Out-of-print books mean the author might control rights again (reversion clauses). Those are easier to license than books still actively published where publishers control rights.

How long does a typical licensing negotiation take?

Trip says every deal has been different. Some close in weeks. Some take months. The platform speeds up the mechanics (verification, payment splitting), but the business negotiation (what rights, what price) still depends on the parties involved.