.png)
AI Writing Is Becoming Commonplace. So Where Does Assistance End and Authorship Begin?
AI Writing Is Becoming Commonplace. So Where Does Assistance End and Authorship Begin?
The argument around artificial intelligence and writing has moved into a more uncomfortable phase. We are no longer only asking whether AI companies should be allowed to train systems on copyrighted books. We are increasingly having to ask what happens when writers themselves use those systems to produce the words that carry their names.
That question sharpened this week when British screenwriter Jack Thorne, president of the Writers’ Guild of Great Britain, warned that some professional writers are using AI to generate entire scripts. His concern is not simply that AI exists in the writing room. It is that undisclosed machine-generated work can be presented as the work of a human writer.
The Guardian reported on 14 September that Thorne wants stronger rules around undisclosed AI-generated professional writing and greater protection for writers whose copyrighted work is used to train models.
Quick Shortcuts
Explore more from 121 Minutes.
The line is becoming more important, not less
There is a temptation to make this debate binary: either you use AI or you do not. I think that is increasingly useless. Modern creative work already sits inside software full of machine-learning features. Research tools, spelling and grammar systems, image tools, transcription, metadata generation and marketing platforms all use forms of automation. The useful question is not whether a writer has ever touched AI. It is what the AI was asked to do.
For me, there is a clear difference between using AI around the creative business of a book and asking AI to be the author of the book. I write my fiction. AI can help me research a subject, visualise a world, test marketing ideas, organise material or create supporting content. Those activities can save time without transferring responsibility for the story itself.
But if a system generates the scene, invents the dialogue, constructs the emotional turn and supplies the finished prose, the question changes. At that point we are no longer discussing a clever tool around authorship. We are discussing how much of the authorship has been delegated.
Publishing is already living with the grey area
This matters because AI is simultaneously becoming routine elsewhere in publishing. Publishers Weekly reported this month that publishers are already using AI in back-office functions such as metadata, royalty statements, proofreading and customer service. That is precisely why a blanket anti-AI position will not solve the problem. The technology can be useful without every use being creatively equivalent.
In its recent industry overview, Publishers Weekly described publishing as still working through where AI helps, where it damages trust and where stronger rules are needed.
For authors, that suggests we need better language. “AI-assisted” should not automatically mean “AI-authored”. Equally, “human-authored” should mean something meaningful. Readers, publishers and fellow writers deserve to know when the creative work they are evaluating was substantially generated by a machine.
The real issue is responsibility
Authorship is not only a claim to credit. It is an acceptance of responsibility. When my name appears on a novel, I am saying that I made the creative decisions, that I stand behind the work and that the successes and failures belong to me.
That is why disclosure matters. I do not think every automated spellcheck or research query requires a warning label. That would make transparency meaningless. But there is a point at which assistance becomes substitution, and professional writing needs to become much better at identifying that point.
The danger is that if we refuse to distinguish between different uses, we create two bad outcomes at once. Responsible writers using AI for legitimate support are treated as though they did not write their books, while genuinely machine-generated writing can hide behind the same vague label of “AI-assisted”. Neither outcome helps authors.
A rule writers can actually use
My working test is simple: if I removed the AI from the process, would the story, argument, characters and prose still fundamentally be mine? If the answer is yes, the system is probably supporting the author. If removing it means there is no meaningful piece of writing left, then calling the result entirely human-authored becomes difficult to defend.
That is not a legal definition, and the industry will eventually need more precise contractual standards. But it is a useful creative one because it puts the writer’s agency back at the centre.
The next stage of the AI debate will therefore be less about whether writers are allowed to use technology and more about what readers and publishers are entitled to know about how a piece of writing was created. That conversation is overdue.
So here is today’s question: where do you draw the line between an author using AI as a tool and AI becoming the author?
About Rob Frankson
Rob Frankson is a science-fiction author and creator of the Near Galaxy Saga. AI & The Author explores the practical, creative and ethical impact of artificial intelligence on writers, authorship and publishing.
Get new AI & The Author articles, author updates and Near Galaxy news from 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.
Comments