British Startup Gets IPO Approval for a Human-Authored Book Seal as AI Anxiety Spreads Through Publishing

Books by People, a British startup, has won approval from the UK Intellectual Property Office to roll out a certification mark for books written by humans, giving publishers a new way to signal that a title was created without AI. The move comes as the publishing world wrestles with a wave of AI-related scandals and growing demand from readers for clearer labeling.

The company’s mark, which resembles a thumbprint, will appear on book covers and is intended to function like an “organic” label for literature: a visible claim that a book was authored by a real person, not a machine. For publishers, the certification is meant to provide reassurance. For readers, it is a promise that the work they are buying was made through a human drafting process.

That promise matters more than ever. In recent months, authors, agents and publishers have faced a string of disputes over whether books were partly or wholly generated using artificial intelligence, raising questions about trust, disclosure and what counts as original authorship in the age of generative AI.

Why human-authored labels are gaining traction

The publishing industry has long assumed that a book on a shelf was created by a person. Until recently, that assumption rarely needed to be defended. AI has changed that, introducing uncertainty about whether a manuscript was written by an author, assisted by software, or produced in substantial part by a model.

Books by People argues that the answer should not be hidden inside technical jargon or buried in fine print. Instead, it believes readers should be able to recognize, at a glance, whether a book was made by a human.

What Books by People says it is trying to solve

Books by People says its aim is simple: preserve reader choice. If someone wants to read machine-generated fiction or non-fiction, the startup says, that is their decision. But if readers want a book that was genuinely written by a person, they should be able to identify it easily.

Esme Dennys, the company’s cofounder, said the organization is trying to create a clear signal for consumers who value human storytelling and want that preference reflected in the marketplace.

Books by People says the label is designed to give readers a straightforward way to choose books that were created by human beings rather than machines, while avoiding a system that assumes every author must prove innocence.

The company’s broader pitch is that publishing should offer transparency without turning authors into suspects by default.

How does the human-authored certification work?

The certification process is more involved than an honor system. Rather than relying only on an author’s promise, Books by People combines declarations, document review and software analysis to decide whether a manuscript passes muster.

According to the startup, authors must sign off on an authorship declaration. The company then reviews early drafts, research notes and related materials supplied by the writer and editor. It also runs manuscripts through proprietary software designed to surface signs that a text may not have followed a normal human drafting process.

The company says it looks for evidence of iterative writing, editing and revision. In other words, it is trying to determine whether a book emerged through the kind of messy, layered process that often characterizes human authorship.

Why the company says it is not chasing AI detection

Books by People insists its process is not simply another race to spot AI-generated text. That distinction matters because detection tools have already become controversial, with critics arguing that they can be inaccurate, opaque and overly confident.

Dennys said the startup recognized early on that trying to match ever-improving AI with ever-more-aggressive detection would not be a durable strategy. Instead, the company says it is placing greater weight on the documentary trail around a manuscript than on the output alone.

When the evidence is ambiguous, the company says it gives the benefit of the doubt to the author rather than risking an unfair accusation.

How the publishing industry got here

The timing of the certification is not accidental. This year alone, publishing has been hit by repeated disputes over possible AI use, and the reputational consequences have been severe.

In one case, a literary agent reportedly cut loose a book deal said to be worth $2 million. In another, a publisher pulled a novel from circulation. Elsewhere, readers questioned whether a prize-winning short story had been produced with AI. Each episode has added to the sense that the industry lacks a settled norm for disclosure.

The result is a growing divide between those who want stronger verification and those who worry that the scrutiny itself may become excessive or unfair.

Why trust is now a business issue

Publishing has always depended on confidence: between editors and authors, publishers and retailers, and the industry and its readers. The fear now is that even isolated cases of AI misuse could erode that trust more broadly.

Some publishers believe that if readers start to suspect machine involvement in books marketed as human-created, the damage will extend beyond individual scandals. The concern is not simply that one title may have been misrepresented, but that the entire category of published literature could be viewed with more suspicion.

Industry figures say publishing has traditionally relied on an informal contract of trust, and that breaches of that contract can quickly feel like fraud to readers who assumed they were buying a human-made work.

Who supports human-only labels, and why?

Support for a human-authored seal has come from writers, editors and some senior publishing figures who see it as a necessary statement of values rather than a perfect enforcement system.

One of the clearest arguments in favor comes from the idea that readers deserve clarity. If a book is marketed as the product of a person’s imagination, labor and judgment, then the industry should be prepared to stand behind that claim.

Marcus Dohle, the former chief executive of Penguin Random House, praised the concept by describing human storytelling as central to the publishing business. His view reflects a wider anxiety that the rise of generative AI could blur the line between authored work and automated content.

What supporters say the seal signals

• That a publisher has made a public commitment to human-created books.
• That readers can identify titles they believe were written without machine assistance.
• That the industry still sees original storytelling as a defining asset.

For some advocates, the label is less about enforcement than principle. Even if no verification method is flawless, they argue, the market should not be left without a visible standard.

What are the risks of AI book detection?

The biggest risk is that detection tools can misfire, and when they do, the consequences can be devastating for authors. A false positive can lead to public suspicion, lost opportunities and lasting damage to a writer’s reputation.

That concern has already surfaced in a highly public way. Jerry Falade, the debut author whose reported book deal was later scrapped, has said detection software misread his Nigerian dialect as evidence of AI use. He said the tool’s verdict arrived with devastating speed, while the writing process itself had taken months or years.

His complaint highlights one of the central tensions in AI policing: software may seem objective, but language is full of regional variation, stylistic quirks and unusual phrasing that can confuse automated systems.

Why false positives worry publishers and authors

False accusations can be especially harmful because the burden of proof often falls on the writer. If a manuscript is flagged, an author may have to explain draft histories, communications and stylistic choices to prove a negative.

Critics say that dynamic can create a chilling effect, making authors wary of experimenting with voice, dialect or unconventional structure. It could also push publishers toward defensive screening practices that treat writers more like potential offenders than creative partners.

Some publishing executives warn that any new verification regime must avoid turning routine authorship checks into a witch hunt, especially if the tools used to flag suspicious text are not reliably accurate.

What do AI detection companies claim?

AI detection has become its own fast-growing niche, with firms positioning themselves as guardians of textual authenticity. One of the best-known is Pangram, a New York–based startup that says it can identify generated text with 99.7% accuracy.

But Pangram’s claims have faced skepticism from authors and critics who question whether tools built with AI themselves can reliably distinguish between machine output and human writing in every context.

That skepticism matters because detection systems are increasingly being asked to do more than merely flag suspicious language. They are being used, directly or indirectly, to influence contracts, publishing decisions and reputations.

How Books by People differs from Pangram

Books by People is not presenting itself as a pure detection startup. Instead, it says its method is broader and more process-oriented, with greater attention paid to manuscript history, drafts and the communication trail surrounding a book.

That approach may be slower, but the company believes it reduces the chance of unfair outcomes. Rather than asking a machine to decide on the basis of text alone, the startup is trying to assess the larger story of how a book came into being.

The difference is important because the publishing world is deciding not just whether AI should be used, but who should have to prove what, and by which standard.

How much will the certification cost?

The startup says publishers will pay a flat membership fee and then an extra amount for each title submitted for certification. That pricing model suggests Books by People is aiming to build a recurring relationship with publishers rather than sell a one-off service.

Five publishers have already signed up, according to the company, and the first labeled book is expected to appear before the end of the year. That early uptake is modest, but it indicates that at least some publishers see value in a visible human-authored marker.

The seal may also become commercially useful if readers begin to seek it out the way some shoppers look for organic food or verified provenance labels. In a crowded marketplace, a stamp that signals “written by a person” could become a differentiator.

Why readers may care more than publishers expect

Many industry observers believe that machine-generated books will never fully replace works written by people, because readers form emotional attachments not only to stories but also to the consciousness behind them.

Ted Chiang, the acclaimed science-fiction author, has argued that the way a work is made affects how it is received. In his view, knowing the origin of a story changes the reading experience itself.

That insight points to the heart of the issue. If a novel is understood as the product of lived experience, revision and human judgment, then the knowledge that it was generated or heavily shaped by AI can alter how readers interpret every sentence.

Chiang has said that the process behind a story cannot be separated from the experience of reading it, because readers inevitably respond to what they know about how the work was made.

What happens next?

The next stage will test whether the certification is more than a symbolic gesture. Books by People still has to prove that its process can scale, that publishers will pay for it and that readers will notice it.

There is also a bigger question: whether the industry can build norms around transparency before a major scandal forces a harsher and less flexible solution. If more publishers adopt human-authored labeling, it could become a new standard. If they do not, the market may remain stuck between overconfident AI detection and underpowered self-reporting.

For now, Books by People is betting that a visible mark of human authorship can restore some clarity to a field being reshaped by generative AI. Its success may depend less on perfect detection than on whether publishers, writers and readers agree that human-made books deserve a label of their own.

Key facts at a glance

• Books by People is a British startup focused on human-authored book certification.
• The UK Intellectual Property Office has approved its certification mark.
• The mark is intended to appear on book covers as a visible sign of human authorship.
• Five publishers have already joined the program.
• The company expects its first certified title to be released by the end of the year.