The Authors’ AI Charter Is Not an Endorsement — Here’s What It Demands

A new international charter on artificial intelligence has landed in the publishing world, and the reaction tells us something important: writers are not simply arguing about technology. They are arguing about trust. Launched on 14 October by the UK Society of Authors alongside fellow organisations in the United States, Australia, Canada and New Zealand, the Charter of Ethical AI Use in the Creative Industries asks publishers and other organisations to make public commitments about generative AI. Yet within a day the Society of Authors felt it necessary to clarify what that actually meant. That clarification matters as much as the launch itself.

In this article

What the charter proposes; why the Society of Authors clarified its position; where AI assistance differs from AI-authored fiction; what authors should ask publishers; and why disclosure must lead to accountability.

A charter, and an immediate clarification

The Society of Authors announced the international initiative on Wednesday, 14 October. Its partners include the Authors Guild in the US, the Australian Society of Authors, the Writers’ Union of Canada, the New Zealand Society of Authors and the UK Creators’ Rights Alliance. The central ambition is to encourage publishing organisations, literary agents and prize bodies to make explicit commitments about how they use generative AI. That might sound uncontroversial. But for authors whose books may have been used in AI training without consent, the language of ethical AI can sound dangerously like acceptance of the technology itself.

On Thursday, chief executive Anna Ganley published a direct response. She said the organisation was not retreating from its criticism of generative AI and did not encourage its use. Rather, the charter was intended to confront the reality that some publishers are already using these systems without sufficient transparency. This is a distinction worth keeping in mind. Establishing standards for something happening in the industry is not necessarily endorsing the practice. It can also be a way of demanding accountability.

What the charter actually asks for

The proposed commitments go beyond a vague promise to behave responsibly. They include assessing risks to intellectual property, data protection, creators' livelihoods and environmental sustainability; developing permitted-use and disclosure policies; training staff; and pursuing consent-based systems. Where AI-generated content materially contributes to published work, transparency is expected. These are organisational commitments rather than a new law, and their value will depend on whether signatories follow through.

There is an obvious test here. If a publisher signs a charter but cannot tell an author whether their manuscript is entered into a generative AI system, the signature is not much protection. If an agent promises transparency but cannot explain which third-party services handle submissions, the author remains in the dark. Policies need to be understandable, enforceable and available before writers are asked to hand over their work.

AI writing a novel is not the same as helping an author run a career

Much of this debate becomes unproductive when every use of AI is treated as identical. It is not. Having a model generate chapters and presenting them as your own imaginative work raises fundamental questions about authorship, originality and reader expectations. Using software to organise research notes, explore possible visual references, prepare a marketing calendar or manage the administrative burden of publishing is a different proposition. Those tasks may still raise questions about privacy, licensing and reliability, but they do not automatically transfer the creative authorship of a novel to a machine.

The boundary is not always neat. An AI-assisted synopsis can become a machine-written synopsis. A research summary may contain fabricated facts. A promotional image may have its own provenance and rights questions. What matters is being specific about the activity, the material supplied to the tool and the degree of human judgement involved. I want technology to reduce the clerical load on authors, not to make their voices interchangeable.

That position does not dismiss writers who object to generative AI altogether. Training practices, energy demands and the economic effect of synthetic content are legitimate concerns. It simply recognises that a professional author may use digital tools without surrendering the imaginative work that makes a book theirs.

Five questions every author should ask

Before signing a publishing agreement, ask whether submitted manuscripts are processed by generative AI, whether any material is retained or used to improve a model, who has access to it, whether the publisher uses AI in editing or publicity, and how an author can challenge an undisclosed use. Ask for answers in writing. The purpose is not to create an administrative obstacle course. It is to establish what happens to your intellectual property once it leaves your desk.

The Authors Guild has also issued guidance this week on publishers' use of AI detectors and undisclosed AI. That is another part of the same trust problem. An unreliable detection score should not be treated as proof of misconduct, and authors deserve clear procedures for contesting accusations. Transparency must work in both directions: authors should be candid about material AI involvement, and publishers should disclose the systems they deploy against authors' work.

A useful beginning, not a finished solution

I see the charter as a practical starting point rather than a resolution. It does not settle the legality of training datasets, recover lost licensing income or prevent every misuse of a manuscript. Nor can a voluntary declaration replace clear contracts or effective copyright enforcement. What it can do is put a simple demand on the table: stop treating AI use inside publishing as something authors need not know about.

The next step is measurable action. Publishers should publish policies that explain permitted uses in plain English, document consent, offer meaningful safeguards and demonstrate that employees actually follow those rules. Authors' organisations should scrutinise implementation rather than count signatures. And writers should be able to make informed decisions about where their work goes.

My question for the industry is straightforward. If generative AI use is responsible, why should it be hidden from the people whose work and livelihoods are affected? A charter becomes valuable when it turns that question into an obligation to answer.

Sources and further reading

Society of Authors — Creative industries launch Ethical AI Charter: https://societyofauthors.org/2026/10/14/creative-industries-launch-ethical-ai-charter/

Anna Ganley — CEO clarification, 15 October: https://societyofauthors.org/2026/10/15/

ceo-blog/

Authors Guild — AI best practices for publishers: https://authorsguild.org/resource/ai-best-practices-for-publishers/

About the author: Rob Frankson writes about the intersection of storytelling, technology and the professional life of an author.

https://www.121minutes.com/about

Join the conversation and subscribe to AI & The Author:

https://www.121minutes.com/subscribe

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.