.png)
AI Licensing Is Becoming Real. What Should Authors Demand?
AI & The Author | 22 September 2026
The AI copyright argument is starting to change shape. After years dominated by scraping, lawsuits and fair-use arguments, publishing is now building something much more practical: a market in which books can be licensed for AI use.
That matters to writers because licensing moves the conversation away from a crude choice between accepting AI or rejecting it. The more useful question is whether authors can retain meaningful control over how their work enters these systems and whether they are paid when it does.
The strongest signal: licensing is becoming real
Publishers' Licensing Services has been arguing that the industry now needs to turn principles into workable licensing solutions. That follows its collective AI licensing scheme, which had attracted more than 250 UK publishers by June. The direction is significant: rather than assuming every book on the internet is simply available for machine training, the publishing industry is developing mechanisms for permission, rights management and remuneration.
At the same time, the $1.5 billion Anthropic copyright settlement has made the value of training material impossible to dismiss as an abstract issue. It has also exposed a difficult secondary question: when money is attached to the use of a book, who is actually entitled to receive it. The author, publisher, agent, or another rights holder? Recent disputes over settlement claims show that authors need to understand their contracts as carefully as the technology.
Why writers should care
For authors, this is potentially a much healthier stage of the AI debate. Litigation will continue, and courts still have major questions to answer about fair use, piracy and training. But licensing asks a different question: if an AI developer wants access to books, what would a fair transaction look like?
My own starting point is straightforward. I write my books. AI can help me research, visualise ideas, organise material, create supporting imagery, market the work and run parts of the creative business around it. None of that makes the AI the author of the fiction.
But the position changes when another company wants to take my finished creative work and use it as an input into a commercial technology. At that point we are no longer talking about my use of AI as a tool. We are talking about somebody else's use of my intellectual property. That is a rights conversation.
AI-assisted is not the same as AI-authored and an author's use of AI is not the same as an AI company's use of the author's book.
What should a fair AI licence contain?
For me, five things matter. First is permission: the author or legitimate rightsholder should know what is being licensed and for what purpose. Second is transparency: there should be a reliable record of which works entered the system. Third is payment: compensation must be meaningful enough to recognise that creative work has commercial value. Fourth is accurate rights data, because publishing contracts frequently divide rights between authors, publishers and territories. Fifth is the ability to say no.
That last point matters. A licensing market is only genuinely voluntary if refusal is a real option rather than a theoretical checkbox attached to a system that already possesses the work.
Publishing is already using AI
This is also why I don't find blanket arguments about whether publishing should 'use AI' particularly helpful. Publishers Weekly recently reported that AI is already being used across publishing for metadata, royalty statements, proofreading and customer service, with 63% of surveyed industry professionals saying their organisations use the technology. AI is becoming part of publishing infrastructure whether the creative community likes every application or not.
The important distinction is what the technology is doing. Automating a royalty statement is not the same creative act as generating a novel. Using AI to visualise a fictional world for marketing is not the same as asking it to invent the characters, scenes and finished prose and then presenting that output as human authorship.
Writers therefore need a more precise vocabulary. We should be able to support useful technology while defending human authorship. We should be able to experiment with AI-assisted creative-business tools while insisting that books have owners and rights. Those positions are not contradictory.
A market could be better than a war
If licensing develops properly, it could create something the current legal battles cannot: a repeatable relationship between authors and AI developers. Developers gain access to high-quality material with clearer provenance. Publishers gain a way of administering rights. Authors gain transparency, choice and potentially a new revenue stream.
There are plenty of ways it could go wrong. Payments could become negligible. Contracts could assign AI rights without authors understanding them. Collective schemes could make opting out difficult. Smaller writers could have less negotiating power than major publishers. Those issues deserve scrutiny now, while the market is still being designed.
But that is precisely why authors should engage with licensing rather than treating it as somebody else's problem. The terms being normalised today could shape how our books are used for years.
The question for writers
We may finally be moving from an era in which creators discover after the fact that their work was used, towards one where permission can come first. That would be progress provided authors remain part of the negotiation rather than merely becoming another line in a rights database.
If an AI company offered to license your novel tomorrow with transparent terms, payment and a genuine right to refuse. What would it have to offer before you said yes?
Quick Shortcuts
Further reading: Publishers' Licensing Services on the emerging AI licensing market; The Bookseller on collective AI licensing in the UK; Reuters on the Anthropic copyright settlement; Publishers Weekly on publishing's wider AI adoption.
About the author
Rob Frankson is a science-fiction author and creator of the Near Galaxy Saga. Through AI & The Author, he explores how artificial intelligence is changing writing, publishing and the creative business — while keeping a clear distinction between technology that supports an author and technology that attempts to replace authorship.
Keep the conversation going
For more AI & The Author briefings, science-fiction, HMSS Broadside and updates from the Near Galaxy, visit 121 Minutes and subscribe for new posts.
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