When AI Suspicion Derails a Book Deal: What Authors and Publishers Need to Consider
A highly anticipated debut novel recently became the center of a growing controversy over artificial intelligence, authorship, and fairness in publishing.
Jerry Falade’s crime novel, Call Me, I’ll Hide the Body, reportedly generated a 14-way auction and an offer exceeding $2 million from Minotaur, an imprint of Macmillan US. The manuscript also attracted international publishing interest and attention from Hollywood.
The momentum abruptly stopped when Falade’s agents withdrew the manuscript after stating that they could no longer authenticate how it had evolved “from origin to completion.” Falade has denied using artificial intelligence and has argued that the response reflects racial bias within the publishing industry.
Regardless of what the underlying facts ultimately prove to be, the controversy highlights several issues that authors, agents, publishers, and film and television buyers can no longer afford to leave unresolved.
What Does It Mean to “Use AI” in Writing?
The first problem is definitional.
There is a considerable difference between asking a generative AI platform to produce chapters of a novel and using technology to assist with research, outlining, grammar, translation, brainstorming, or editorial feedback. Yet contracts and submission guidelines often refer broadly to “AI use” without identifying which activities are prohibited or must be disclosed.
That ambiguity creates risk for everyone involved. An author may reasonably believe that using an AI-enabled editing feature is no different from using conventional spelling or grammar software, while a publisher may view the same use as a material failure to disclose.
Publishers should clearly define:
What qualifies as generative AI;
Whether any AI-assisted activity is permitted;
Which uses must be disclosed;
Whether disclosure is required at submission, acquisition, or delivery;
Whether different rules apply to generating expressive content and performing administrative or editorial tasks; and
What documentation the author must maintain.
Authors, in turn, should not assume that a publisher’s silence means all forms of AI assistance are acceptable.
Proving How a Manuscript Was Created
Historically, publishers have largely relied on an author’s representations that the manuscript is original and does not infringe another party’s rights. Generative AI complicates that trust-based system.
Authors may increasingly be asked to demonstrate how their work developed. Drafts, handwritten notes, research files, tracked changes, cloud-based version histories, correspondence with editors, and dated outlines can help establish the evolution of a manuscript.
This does not mean writers should be required to prove their innocence whenever an allegation arises. It does mean that maintaining organized records of the creative process is becoming an important part of protecting a literary work much like preserving chain-of-title documentation is essential when adapting a book for film or television.
Publishers should also establish in advance what evidence will be considered sufficient. Requiring proof without defining the standard can lead to subjective and inconsistent decisions.
AI Detectors Are Not Conclusive Evidence
AI-detection tools may appear to offer a quick solution, but their results should be treated cautiously. These systems generally estimate whether language resembles machine-generated text; they do not have direct knowledge of who wrote a manuscript or how it was created.
False positives can have severe consequences. A writer could lose representation, a publishing agreement, adaptation opportunities, and professional standing based on an assessment that cannot definitively establish AI use.
Any meaningful review process should consider multiple forms of evidence rather than relying on a detector score alone. It should also give the author notice of the specific concerns and a reasonable opportunity to respond before irreversible decisions or public statements are made.
AI has created a genuine verification problem, but suspicion is not the same as proof.
Copyright and Chain-of-Title Concerns
For publishers and Hollywood buyers, the central legal question is whether they are acquiring enforceable rights.
Under current U.S. Copyright Office guidance, copyright protects human-authored expression. Material generated entirely by AI may not qualify for copyright protection, although sufficiently creative human-authored selection, arrangement, revision, or modification may still be protectable. The analysis is highly fact-specific.
The mere use of an AI tool does not necessarily make an entire manuscript unprotectable. The relevant questions include what the tool produced, how much of that output remains in the final work, and the extent of the author’s creative control and contribution.
Nevertheless, undisclosed AI-generated material can create uncertainty in the chain of title. A publisher acquiring literary rights and a studio later acquiring adaptation rights needs confidence that the author owns the material being licensed or assigned. If portions of the work are not copyrightable, or if the work incorporates potentially infringing material, the value and exclusivity of those rights may be affected.
These concerns are likely to lead to more detailed contractual representations, warranties, disclosure requirements, indemnities, audit rights, and termination provisions.
Contract Language Must Be More Precise
A blanket warranty that a manuscript was created “without AI” may be too vague to serve either party well.
An effective clause should address:
The specific technologies and activities covered;
Whether AI-generated expressive content is prohibited;
Whether limited editorial or research assistance is permitted;
The author’s disclosure obligations;
Whether prior written approval is required;
The author’s obligation to preserve supporting records;
Procedures for investigating suspected violations;
The opportunity to explain or cure a potential breach;
Confidentiality during an investigation; and
The remedies available if a representation proves inaccurate.
The agreement should also distinguish between an intentional misrepresentation and an inadvertent or immaterial use of an AI-enabled tool. Immediate termination may be appropriate in some circumstances, but it should not automatically be the only available response.
Fairness, Bias, and Reputational Harm
Falade has asserted that Black authors receiving significant publishing attention have been subjected to disproportionate suspicion about whether their work was truly their own. That allegation deserves serious consideration.
When an exceptional manuscript is treated as evidence that the writer may not have created it, questions arise about whose talent is readily believed and whose authorship must be authenticated. Even facially neutral review procedures can produce unfair results if they are applied inconsistently or influenced by unconscious bias.
Publishers and agencies should adopt uniform procedures for investigating AI concerns. The same evidentiary standards, response periods, and decision-making processes should apply regardless of an author’s identity, background, commercial success, or level of industry access.
The stakes extend beyond the immediate contract. Public allegations concerning authorship can cause lasting reputational harm even when AI use is never established. Publishers and representatives should therefore carefully consider when an internal contractual concern becomes a matter for public disclosure.
A Need for Industry-Wide Standards
Publishing is confronting an authentication problem without a reliable authentication system.
Clearer contractual language is one part of the solution, but the industry also needs consistent terminology, responsible investigation procedures, limits on the use of detection software, and meaningful opportunities for authors to respond. Without those safeguards, publishers face the risk of acquiring works with uncertain rights, while authors face the equally serious risk of losing careers based on unverified allegations.
Generative AI is changing the creative process, but it should not eliminate fundamental principles of fairness. The goal should be to protect human authorship, copyright ownership, and contractual trust without turning every successful or unconventional work into a cause for suspicion.
Authors and publishers should address these issues before a manuscript is submitted or acquired—not after questions arise and valuable deals are placed in jeopardy.
Legacy Arts Law assists authors and publishers with drafting and negotiating AI contract clauses, authorship representations and warranties, disclosure provisions, publishing agreements, and related chain-of-title concerns. Contact Legacy Arts Law for assistance developing clear, practical AI provisions tailored to your publishing relationship.
This article is provided for general informational purposes only and does not constitute legal advice or create an attorney-client relationship.