AI Publishing's Proof Problem: Rights, Detectors and Author Identity

The week publishing began asking for proof

This week's most important developments share a theme: publishing needs evidence and accountability, not merely declarations of good intentions. A new ethical charter, fresh guidance on AI detectors, an expanding content-rights coalition and reports of author impersonation all ask who controls creative work and who must prove how it is used.

The change is significant because the pressure is moving beyond the manuscript. Authors, publishers, technology companies and online platforms are all being asked to explain their decisions. For an independent writer, the stakes are not abstract. They involve rights, contracts, reputation, discoverability and whether readers can trust the name on a book.

Quick Shortcuts

Ethical AI rules for publishers · AI detectors and fair evidence · Licensing and measurable content use · Author identity and fraud

Ethical AI rules: transparency must run both ways

On 14 October, the UK Society of Authors and international writers' organisations launched a Charter of Ethical AI Use in the Creative Industries. Its principles centre on consent, licensing, ethical use, accountability, rights and remuneration. It asks publishers and other organisations to disclose policies and assess risks before using generative AI in creative workflows. That matters because the people who commission and distribute books should not operate under looser standards than the people who write them.

The next day, Society of Authors chief executive Anna Ganley clarified that the charter is not an endorsement of generative AI. The organisation continues to warn about damage to writers' livelihoods and calls for effective protection. A voluntary charter cannot recover a lost royalty or enforce a contract, but it can establish a benchmark against which the industry can be judged.

As a working author, I want more than a logo on a policy page. I want to know whether a manuscript will be uploaded to an AI service, whether the system retains it, who can approve that use and whether a publisher will disclose generated cover or marketing material. Clear permission and a written policy are the beginning of a professional relationship, not unreasonable demands.

AI detectors: suspicion is not a verdict

The Authors Guild issued new best practices for publishers on 13 October. It urged written AI policies, transparency about detector use and author permission before manuscripts are uploaded to third-party detection services. Most importantly, the Guild warned that AI detector scores alone are not reliable enough to justify accusing an author or cancelling a contract. Its guidance also rejects mandatory monitoring of every writer's drafting process.

The human consequences were visible in a continuing Canadian publishing controversy. On 15 October, Canadian Press reported that Quebec publisher Éditions du Boréal would keep Thélyson Orélien's novel on sale despite unresolved questions about its creation. Orélien denies using AI to write the novel, although he acknowledged plagiarism in earlier work. Those facts must not be conflated: a separate admission does not establish that a novel was machine-written.

My view is that writers should keep drafts, notes and revision histories because they are useful for their own craft and business. But a sensible record is different from requiring every author to prove innocence to an algorithm. When there is a genuine concern, editors should investigate, talk to the writer and use clear contractual standards. A machine's probability estimate should never be the entire case.

Copyright: measurement may become as important as permission

On 12 October, more than 40 news publishers, including The New York Times, Agence France-Presse and The Economist, joined the Standards for Publisher Usage Rights coalition. SPUR now has more than 80 members across over 20 countries. It is developing technical standards for recording AI use of publisher content and supporting licensing negotiations.

This is a media-publishing initiative, not an established solution for novelists. Nevertheless, its logic is relevant to books. How can a writer negotiate compensation when neither the writer nor the publisher can tell whether, when or how the work was used? Transparent reporting and consent-based licensing could eventually make rights meaningful in practice. The unresolved questions are adoption, verification and enforcement.

There is a second commercial implication. If AI-generated answers satisfy readers without sending them to original sites, creators lose opportunities to earn attention, subscriptions and sales. For independent authors, a trusted website, newsletter and recognisable identity are therefore not optional decoration. They are increasingly part of the business infrastructure.

Author identity is becoming a security issue

On 17 October, journalist Linda Geddes described an extraordinary impersonation case: people using established writers' names to pitch articles, with at least one fraudulent byline reaching publication. Her investigation also reported books wrongly attributed to a medical author. Some pitches bore signs of AI assistance, although the perpetrators and their motives remain uncertain.

That is a warning for novelists as well as journalists. Reputation takes years to build. An impostor's email, fake profile or misleading book listing can damage relationships with readers, publishers and agents. A consistent official author site, verified contact channels and regular checks of retail listings offer practical protection. Unexpected publicity offers should be independently checked, however flattering they sound.

Where I stand as a science-fiction author

I write science fiction because I want to create stories, worlds and characters that are mine. The imaginative choices and the writing itself remain the centre of that work. I do not regard asking a machine to produce a novel as equivalent to using technology to research an idea, visualise a setting, maintain continuity, prepare promotional material or manage the business surrounding a book.

That distinction is important but not an excuse to ignore difficult questions. AI-assisted images and marketing can still involve disputed training data, consent and disclosure. Writers must take responsibility for the tools they choose. Equally, publishers should not be allowed to demand strict transparency from authors while keeping their own AI workflows hidden. Trust must be reciprocal.

What writers should watch next

The next test is implementation. Will charter signatories publish clear, enforceable policies? Will publishers revise contracts to specify detector procedures and manuscript safeguards? Will rights standards produce verifiable records of AI use? Will retailers and commissioning editors improve identity checks without unfairly burdening legitimate authors? Those outcomes matter more than another week of declarations.

For now, I would review AI clauses in contracts, preserve the working history of active manuscripts, check official book listings and keep direct reader channels healthy. None of these measures settles the wider legal disputes. They do, however, help protect the creative work and the professional identity behind it.

The question I am left with

If publishers want authors to prove their work is human, should publishers first have to prove that their own use of AI is disclosed, consent-based and accountable?

Sources and further reading

- Society of Authors --- Ethical AI Charter, 14 October • Society of Authors --- CEO clarification, 15 October • Authors Guild --- AI Best Practices for Publishers, 13 October • Reuters --- SPUR publisher coalition expansion, 12 October • Canadian Press --- Orélien publisher statement, 15 October • The Guardian --- Linda Geddes on author impersonation, 17 October

About Rob Frankson

Rob Frankson is an independent science-fiction author behind 121 Minutes and the Near Galaxy Saga. He explores original storytelling, emerging technology and the practical business of being an author. About Rob Frankson.

Join the conversation

Subscribe for more AI & The Author analysis and updates from 121 Minutes. Join / Subscribe to 121 Minutes.

AI & Editorial Transparency

AI & The Author is edited and published by Rob Frankson. Artificial intelligence is used to assist with news research, initial drafting, content organisation and supporting imagery. All articles are reviewed and, where necessary, edited by Rob Frankson before publication. The opinions, editorial position and final decision to publish remain the author's.