You have no items in your cart. Want to get some nice things?
Go shoppingWhere Literature Happens · Frankfurt
Who Owns the Future of Publishing?
At Frankfurt Book Fair, the business of books is being reshaped by artificial intelligence, translation, international rights and a new struggle over the value of published work. Eric Akoto reports from inside the industry.
A book can begin in one country, acquire readers in another and become a commercial proposition in a dozen more. Somewhere between those possibilities lies the business of publishing: the negotiation of rights, the cost of translation, the judgement of editors and the difficult calculation of whether an unfamiliar story will find an audience.
At Frankfurt Book Fair, those decisions are being made in meeting rooms, at publishers’ stands and in conversations that may last only a few minutes.
But the questions facing the industry are becoming larger than which books will sell next season.
Artificial intelligence has introduced new possibilities for translation, discovery and distribution. It has also created an entirely different commercial problem. Companies developing AI systems want access to enormous quantities of published material. Publishers, meanwhile, are trying to determine what that access should cost, which uses they can authorise and whether the value of their work is being transferred elsewhere.
The resulting arguments reach far beyond technology.
They concern the future of literary ownership, the economics of bringing books across borders and the role publishers will occupy if machines can perform some of the work that once distinguished them.
The cost of crossing borders
At a Frankfurt rights gathering, I met Monika Drummer, who was entering the business of foreign-rights representation for German writers.
She was studying the movement of genres between markets, including the relationship between American publishing trends and German readerships. Together with her husband, she was developing a data tool to help identify patterns and potential opportunities.
Her interest extended to Asian fiction and the possibility of finding new audiences for writing that had not yet travelled widely.
This was an instructive conversation because it brought the international publishing business down to the scale of a single agent making decisions about where to invest her attention.
A book’s success in its home market is no guarantee that it will succeed elsewhere. Its genre may not align with local buying habits. The publisher may be reluctant to absorb translation costs. An editor may admire the writing but doubt that sufficient readers can be found.
These are not new problems. They have shaped international publishing for decades.
What is changing is the quantity of information available to inform those judgements — and the growing possibility that some of the work involved in crossing language barriers will become less expensive.
Data can identify patterns. Technology can accelerate parts of the process. But somebody still has to decide whether a book is worth bringing to a new readership.
That distinction matters.
Literature does not travel simply because the words can be translated. It travels when publishers, agents, translators and readers create the conditions in which a work can be understood and valued.
The commercial question is whether emerging technologies will widen those opportunities or concentrate more influence among those already controlling the market.
When rights become the product
Publishing has always depended upon the management of rights.
An author grants a publisher permission to exploit particular rights under agreed conditions. Foreign editions, translations, audio adaptations, film and television projects can generate additional commercial lives for a work.
These arrangements are rarely simple. Rights may be limited by territory, language, format, duration or purpose. The same book can support several distinct business relationships without any single party controlling every possible use.
Artificial intelligence makes those distinctions more important, not less.
During a Frankfurt discussion of copyright and licensing, Roy Kaufman of Copyright Clearance Center argued for an active commercial response.
His instruction to publishers was to get out there and license.
It was an unusually direct formulation of a position increasingly visible within the industry: rather than treating every use of publisher content by an AI company as an unavoidable technological development, publishers should establish the terms under which their material can be used.
But what exactly is being licensed?
There is a significant difference between providing material to train a model and allowing an authorised system to retrieve passages or information from a collection of publications.
Training may involve the use of works in developing a model’s capabilities. Retrieval-based systems can instead consult identified sources while responding to a particular request. Other AI applications may perform narrower functions within research, editorial production or professional information services.
These uses raise different questions about permission, attribution, compensation and control.
Treating them as one activity risks obscuring where value is created and where rights may be lost.
For independent publishers, this is particularly consequential. They may possess valuable archives but lack the negotiating resources, technical infrastructure or legal budgets available to larger publishing organisations.
The existence of a potential licensing market does not guarantee that every publisher will be able to participate on equal terms.
The question of value
Caroline Sutton, representing the interests of scholarly publishing through STM, raised another important question about the relationship between publishers and AI platforms.
Her position was that publishers should make authorised content valuable enough for AI companies to seek legitimate access rather than rely on unauthorised collection.
That sounds commercially sensible. Yet it leaves a difficult question unanswered.
What constitutes the value of a published work in this new market?
Is it the reliability of the information? The editorial investment made before publication? The integrity of the archive? Its specialised subject matter? The right to reproduce or process the underlying content?
Or is its value determined mainly by the capacity of technology companies to use it at scale?
For scholarly and professional publishers, the answer may involve established research workflows, authoritative records and the continuing needs of institutions that depend on accurate information.
For literary publishers, the calculation is less straightforward.
A short story, poem or novel does not necessarily possess value because it supplies a factual answer. Its originality may reside in voice, form, imagination or the particular relationship it establishes with a reader.
Those qualities are central to literature. Whether they translate into a sustainable AI-licensing market is a separate question.
This is where the debate becomes uncomfortable.
A technology may find published material useful without necessarily preserving the economic conditions required to produce the next generation of that material.
Experiment, but on whose terms?
Chris Reid of Wiley discussed the importance of experimentation and understanding how researchers and users are changing their behaviour.
His contribution pointed towards a broader transformation in publishing: the business increasingly involves not only supplying content but understanding how people work with it.
For research publishers, that might mean new tools, services and forms of access developed around existing collections.
For trade publishers, the equivalent opportunities are less settled.
There may be improvements in translation, discoverability, editorial workflows and rights administration. But efficiencies in production do not automatically lead to greater diversity in what gets published.
Lower costs can create opportunities for smaller organisations. They can also make it easier to produce more material than the market can reasonably absorb.
The question is therefore not simply how publishers can adopt AI.
It is which activities they should continue to control, which they can responsibly automate and which remain essential to the cultural identity of their businesses.
Mehmet Toral of Frontiers raised the complications surrounding Creative Commons licensing and AI training. The availability of openly licensed material does not necessarily remove every question about particular downstream uses.
The phrase that kept emerging from this part of the discussion was familiar from copyright debates: it depends.
It depends on the licence, the jurisdiction, the activity being undertaken and the rights involved.
That uncertainty is commercially inconvenient. It is also a reason to resist the idea that one technology, licence or industry agreement will resolve the entire relationship between publishing and AI.
The publisher’s judgement
For all the discussion of data, licensing and technology, Frankfurt remains a place where somebody has to decide whether a book deserves attention.
That decision may be informed by sales records, an author’s readership, market trends or the enthusiasm of an editor. It is still a judgement about an uncertain future.
A rights agent studying international demand, an academic publisher negotiating AI access and a small literary press looking for its next readership occupy very different positions in the industry.
But each faces a version of the same problem: how to establish the value of work before its future value is known.
The temptation is to think that better data will gradually eliminate this uncertainty.
It may reduce some of it. The translation of a manuscript can become faster. Market information can become more accessible. Licensing can be made more systematic.
None of those developments guarantees that the right books will be recognised, that authors will be treated fairly or that publishers will discover work whose value cannot yet be measured.
Publishing is an unusual business because its products are also cultural works. A book may sell poorly on publication and become important years later. A translation may find an audience the original publisher never imagined. A small literary magazine may publish a writer long before there is any obvious commercial interest in their work.
Those possibilities depend upon institutions and individuals willing to make judgements that cannot be reduced entirely to short-term returns.
Frankfurt’s conversations about the future of publishing therefore lead back to an older question.
Who is prepared to invest in literature when its value is not immediately apparent?
Artificial intelligence may change how books are discovered, translated, licensed and used. It will undoubtedly change parts of the economics surrounding them.
But a publishing industry organised solely around extracting value from existing works risks neglecting the more difficult activity on which its future depends: finding, developing and supporting work that does not yet have a market.
That is the business of publishing worth defending.
Continue reading
What Is a Literary Festival Doing in Magaluf?
Dispatch 01 of Where Literature Happens: at FLEM 2026 in Mallorca, writers, readers and a hotel test the idea that literature belongs wherever people gather.



