Skip to main content
Se Habla EspaƱol
|
|(210) 742-2435
Back to Resources

Copyright

AI and Copyright Law: What a $2.4 Million Book Deal Falling Apart Teaches Us

A debut novelist's dream book deal collapsed within hours, not over the quality of the writing, but over a question nobody could answer with certainty: how much of the manuscript was written with the help of artificial intelligence.

In mid-2026, the publishing world watched a debut novelist's dream deal collapse within hours. Jerry Falade's crime novel, Call Me, I'll Hide the Body, had sold to Minotaur Books, an imprint of Macmillan, in a deal reportedly worth more than two million dollars for that book and a follow-up, after a fourteen-way bidding war. Then, before publication, the deal was withdrawn. The reason had nothing to do with the quality of the writing. It had to do with a question nobody could answer with certainty: how much of the manuscript, if any, was written with the help of artificial intelligence.

Falade has maintained throughout that he wrote the book himself. His agents, in a message to publishers, said they had trusted his assurances during the submission process but could no longer authenticate how the manuscript had evolved from start to finish. Whether AI was actually involved remains genuinely unresolved. What is not in dispute is the legal principle that made publishers this cautious in the first place, and it is worth understanding if you create anything, whether that is a novel, a screenplay, marketing copy, or a piece of software.

The Rule: Copyright Protects Human Authors

The United States Copyright Office has been consistent on this point across multiple rounds of guidance, most recently a detailed report issued in January 2025. Human authorship is a requirement for copyright protection, not a preference. A work created entirely by an AI system, with no meaningful human creative control over the expressive result, does not qualify for copyright at all. This principle was tested directly in federal court in a case involving a fully AI-generated image, and the courts sided with the Copyright Office's position.

This is not simply a policy the Office invented. It reflects a basic structural feature of copyright law: it protects the expression of an idea, created by a person, not the output of a process.

Why "I Used AI a Little" Is a Harder Question Than It Sounds

Most real-world situations are not as clean as entirely AI or entirely human. The Copyright Office's guidance addresses this directly. When a work contains both human and AI-generated material, only the human-authored parts are potentially protectable. Using AI as a tool, the way you might use a word processor's grammar checker or a search engine for research, does not strip copyright from the human creative work built around it. But if AI is used to generate substantial expressive content, such as full scenes, sentences, or passages, that specific material generally cannot be copyrighted, even if a human edited it afterward, unless the human's changes themselves rise to the level of new authorship.

The Office has also been clear that prompts alone, no matter how detailed, do not make the person who wrote the prompt the author of what the AI produces. Typing a very specific instruction into an AI tool is different, legally, from writing the sentences yourself.

Why This Became a Business Problem, Not Just an Ethics Question

This is the part that actually sank the Falade deal. A publisher is not just buying a story. It is buying the exclusive right to reproduce, distribute, and adapt it, and it needs to know that right is real and enforceable. If a meaningful portion of a manuscript turns out to be AI-generated, that portion is not protected by copyright at all, which means, in theory, anyone could legally copy it. A publisher who has just paid millions of dollars, and a film studio negotiating adaptation rights on top of that, has good reason not to gamble on an unclear title chain. Reporting on the Falade situation described exactly this concern: not necessarily that AI use is inherently wrong, but that its mere possibility made the ownership of the work impossible to guarantee.

What This Means If You Create Anything

You do not need to be negotiating a seven-figure book deal for this to matter. If you use AI tools in your work, whether you are a writer, a marketer, a small business owner producing content, or a software developer, it is worth understanding which parts of what you make are actually protectable and which are not. Keep records of your creative process. If you are working with a collaborator, an agent, or a publisher, be precise and honest about where AI fit into the process, since vague or shifting answers create exactly the kind of doubt that ended Falade's deal, regardless of what actually happened. And if AI played a meaningful role in generating content you plan to register for copyright, get advice on what you can accurately disclose and what you can still claim as your own.

The technology is moving faster than most people's understanding of what it does to their legal rights. The Falade situation is a visible, expensive example of a question that comes up quietly, every day, for far smaller stakes: when you create something with the help of AI, how much of it is actually yours?