Jerry Falade had the kind of publishing debut that makes other writers stare resentfully at their laptops. His crime novel, “Call Me, I’ll Hide the Body,” sparked a 14-publisher auction. Minotaur reportedly agreed to pay more than $2 million for U.S. rights to the novel and a second book. U.K. rights went to HarperCollins. Translation and screen opportunities were circling.

Then the whole thing collapsed.

Falade’s literary representatives withdrew the manuscript after saying they could no longer authenticate how it had evolved from beginning to end. Falade denies using AI to write the novel. He has acknowledged using an open-source AI tool for research, including delightfully crime-novel-ish questions about disposing of bodies, but says the prose itself was his.

The public still does not know what caused the final break in trust. That matters. The available reporting does not identify an AI detector score, incriminating prompt history, suspicious metadata or any other tidy piece of evidence that conclusively proves AI wrote the book.

What we do know is stranger. An editor reportedly raised an AI concern while the manuscript was being shopped. Falade’s agents investigated, spoke with him and initially found his explanation credible. They did not rely on AI detection software. The auction continued. Only later, after another meeting in which the agents say aspects of Falade’s account changed, did they pull the book.

That chronology should make every creative professional pay attention.

Follow the Money

There is a detail here that is easy to miss when the story gets shortened to “publisher kills $2 million book deal.” Falade’s agents withdrew the manuscript themselves.

That is a very expensive act of skepticism.

A literary agent typically earns about 15 percent. On a $2.4 million deal, that could mean roughly $360,000 in commission before adding U.K., translation, film, television and future-book opportunities. These were people with every financial incentive to keep the train moving, and they pulled the emergency brake.

That does not prove Falade used AI. It does suggest the agents believed they could no longer stand comfortably behind the story of how the manuscript had been created.

Their wording is revealing. They did not publicly declare that Falade had used AI to write the novel. They said they could no longer authenticate the manuscript’s evolution from origin to completion. In the world of rights deals, that distinction is enormous. An agent selling a valuable intellectual property asset needs confidence that the seller can make the representations a buyer will expect.

Once that confidence disappears, the deal can become radioactive even without courtroom-grade proof of misconduct.

So How Do You Prove a Negative?

Here is the problem for every writer, screenwriter, musician, photographer, designer and illustrator watching this unfold: how exactly do you prove that you did not use generative AI to create something?

Short answer: You probably can’t prove it with absolute certainty.

A finished Word file does not contain a little certificate saying, “Congratulations, this paragraph came directly from a human brain.” AI detectors do not fix the problem. They can miss machine-generated text, flag human writing and be defeated by editing. A detector may be one clue in an investigation. Treating it as a verdict is asking for trouble.

The stronger answer is provenance. Creators may need to preserve a record of how the work developed, almost like a chain of custody for creativity.

For a writer, that could mean version history from Google Docs or Word, dated backups, outlines, handwritten notes, character sketches, research folders, discarded scenes, tracked changes and emails with editors or early readers. None proves human authorship on its own. Together they can tell a persuasive story of a manuscript growing over time.

The same idea carries across creative fields. A photographer can preserve native files. A designer can retain layered project files and sketches. A musician can keep demo recordings and session histories. A filmmaker can preserve edit timelines, source footage and earlier cuts.

The evidence becomes much more credible when it exists before anyone asks for it. A folder assembled after an accusation will always look different from months of ordinary creative debris that accumulated naturally while the work was being made.

Creative People May Need Better Bookkeeping

That is a peculiar cultural shift. Creative professionals have traditionally been judged on the finished work. The next era may require them to keep receipts for the creative process too.

This does not mean every novelist should preserve a forensic archive worthy of the Manhattan District Attorney. It does mean basic recordkeeping may become part of professional hygiene, especially when substantial money or valuable rights are involved.

Creators should also get more precise when talking about AI use. “I did not use AI” is becoming almost useless as a statement. Did you use it for search? Research? Brainstorming? Grammar suggestions? Translation? Rewriting? Generating dialogue? Producing final images? Cleaning audio? The tools are becoming embedded in ordinary software, so blanket declarations will get harder to interpret and harder to enforce.

Contracts will likely get more specific too. Publishers, studios, labels and agencies can be expected to distinguish between permitted research assistance, disclosable generative help and uses that affect authorship, ownership or the value being purchased. The important questions will concern who made the expressive choices, what the machine contributed, whether that contribution was disclosed and whether the seller owns what the buyer thinks it is buying.

The Risk of an AI Witch Hunt

There is another side to this, and it may be just as important.

Once suspicion of AI use can destroy a career, suspicion itself becomes dangerous. A creator could be punished because an editor thinks the prose feels synthetic, because a detector spits out a scary percentage or because the creator cannot produce years of pristine documentation on demand.

That burden will not fall evenly. Established writers have editors, agents, old drafts, correspondence and a long public record of voice. Debut writers may have little more than the manuscript sitting in front of them. Newcomers can therefore face the greatest demand for proof while having the least institutional protection.

There are also privacy concerns. A writer trying to defend authorship might be pressured to hand over drafts, account histories, device metadata, research logs and private correspondence. The industry should be careful not to solve secret AI use by normalizing invasive surveillance of creators.

And false accusations will be hard to undo. A novelist who becomes known as “the author accused of using AI” may carry that label even if no one ever proves anything. In creative industries, reputation is part of the asset.

The New Premium: Verifiable Humanity

AI can already produce enormous volumes of acceptable creative material at almost no marginal cost. That creates a strange two-front problem for humans. They may have to compete against a flood of inexpensive synthetic content while also proving that their own work is not synthetic.

That could make verified human creation more valuable. We may eventually see a market for “human-made” work much the way consumers now respond to labels such as handmade, live or organic. The label will mean very little unless there are sensible standards behind it.

The goal should not be to pretend AI has no legitimate place in creative work. It plainly does. Research, ideation, editing, visual effects, accessibility, translation and countless other uses will become ordinary. The industry needs rules that separate assistance from authorship in ways people can actually understand.

A workable system would require clear disclosure standards, reasonable record retention, qualified human review and a meaningful opportunity for the creator to answer specific concerns. Detector output should never be enough by itself to sink a major deal. Accusations should have consequences when they are reckless, just as deliberate misrepresentations should have consequences when they are proven.

The Technotainment Takeaway

The Falade controversy may eventually produce a clean explanation. Perhaps there is powerful evidence we have not seen. Perhaps the decisive issue was simply that his agents no longer trusted changing answers. Perhaps litigation or a fuller public response will tell us exactly what happened.

For now, the uncertainty is the point.

A creator can produce a work that buyers love, trigger a multimillion-dollar auction and still face a second test: can the creator establish a credible history of human authorship?

That is a remarkable new burden. Talent may get you in the room. Process may increasingly determine whether the deal closes.

Creative people have spent years being told to save their drafts because they might want them someday. Turns out “someday” may now include the moment someone asks you to prove that you actually made the thing they are trying to buy.