.png)
AI, Copyright and the Human Author: Publishing Draws New Lines
This week the argument about AI and publishing moved another step away from theory. The questions are becoming practical: what counts as fair use, how should platforms respond to industrial-scale AI publishing, and how do writers protect the value of human authorship without pretending that every use of artificial intelligence is the same thing?
The legal ground is beginning to harden
The most consequential development came from the United States. On 29 September, a federal appeals court upheld Thomson Reuters’ copyright victory over Ross Intelligence. The case involved Westlaw headnotes used to build a competing legal AI search product, and the court rejected Ross’s fair-use defence. It is not a blanket ruling on generative AI, but it matters because an appellate court has now reinforced the principle that commercially valuable copyrighted material cannot simply be treated as free raw material when it is being used to create a competing product.
For authors, that distinction matters. The copyright argument is no longer just an abstract fight between writers and technology companies. Courts are starting to define where commercial competition, transformation and permission intersect. Writers should watch carefully because those principles will influence the much larger disputes over books used in model training.
Platforms are being forced to deal with scale
The second important shift is happening in self-publishing. Amazon has tightened Kindle Direct Publishing upload limits to two manuscripts per week per format. The change arrives amid mounting concern about AI-generated books flooding digital marketplaces. A limit cannot determine whether a book is good, human-written or machine-generated, but it can make mass uploading harder.
That is significant for legitimate independent authors. Self-publishing was built around removing gatekeepers; generative AI has made it possible to exploit that openness at extraordinary scale. The danger is that platforms respond with increasingly blunt restrictions that inconvenience genuine writers as well as content farms. The challenge now is to suppress industrialised spam without rebuilding the barriers independent publishing spent years dismantling.
Publishing is moving toward disclosure rather than prohibition
Another signal came from South Korea, where an industry survey reported that 76% of publishers already use generative AI somewhere in their production processes, while discussion has begun around an AI-involvement label for books. That is a more useful conversation than simply asking whether AI was used. Publishing workflows already contain many forms of automation; what readers increasingly need to know is where the creative authorship actually sits.
For me, this is the distinction that matters. I do not see AI-generated fiction and an author using AI to research a subject, visualise a fictional world, prepare marketing material or organise the business around a book as equivalent activities. One delegates the central creative act. The other can support it. A useful disclosure system needs enough nuance to recognise that difference.
Human review is becoming part of the AI workflow
That distinction also matters outside books. On 2 October, Google guidance highlighted the need for publishers to manually fact-check AI-assisted material before it goes live, including metadata and supporting content. Whatever tools an author uses, responsibility cannot be automated away. If my name is on an article or a novel, I remain responsible for what reaches the reader.
That principle is likely to become increasingly valuable. As generated material becomes cheaper and more plentiful, editorial judgement, provenance and trust become differentiators rather than old-fashioned overheads.
The UK creative argument is changing too
The UK debate over copyright remains important internationally. Creative-industry resistance to proposals that would have made it easier for AI companies to use copyrighted work without permission has become a warning to policymakers elsewhere. The political lesson is increasingly clear: creators are willing to accept new technology, but that does not mean accepting the idea that creative work has no economic value once a machine wants to learn from it.
What authors should watch next
Three things now deserve attention. First, watch the next copyright decisions rather than assuming any single case settles the training question. Second, watch how Amazon and other marketplaces refine their response to AI-generated volume; platform rules may affect independent authors faster than legislation does. Third, watch disclosure standards. A simple AI/no-AI label is unlikely to be sufficient for a creative workflow that may involve research, editing, translation, visualisation, marketing and generation in very different proportions.
My view as a working author
The strongest position for authors is not to become anti-technology. It is to become clearer about authorship. The story, the choices, the voice and the responsibility for the finished work should remain recognisably human. Around that core, technology can be extraordinarily useful. I use that distinction because it lets us defend the value of writing without pretending the tools around writing must remain frozen in time.
The more synthetic content fills the market, the more valuable genuine creative identity may become. Readers do not only buy words arranged on pages. They buy judgement, imagination, experience and a relationship with the person whose name is on the cover.
The question for writers
If publishing moves toward formal AI disclosure, what should readers actually be told: simply whether AI was used at all, or exactly where the boundary between human authorship and machine assistance was drawn?
About Rob Frankson
Rob Frankson is the author behind 121 Minutes and the Near Galaxy Saga, writing original science fiction while exploring how emerging technology is changing the practical business of being an author. About Rob Frankson
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