Innovations in Publishing Symposium 2026

Photo credit: © Oswin Tickler

In September 2026, we hosted our sixth symposium, welcoming another group of researchers to discuss innovations in publishing. The researchers presented a variety of work exploring new approaches to publishing alongside current developments across various segments of the publishing landscape.

Topics include the impact and use of AI on the publishing industry in Europe, the construction of authorship, an overview of the platformised webnovel industry, the use of programming for print making, the graphic in translation and translation dynamics of globally successful fiction.

Rewriting’ the Webnovel Industry: The Convergence of Free Content and AI

Dr Xiang Ren, The University of Sydney, Australia

Xiang presented research on disruptive changes in the digital publishing industry, driven by the convergence of free content models and AI integration. Focusing on the rapid growth of China’s webnovel industry, also known as online literature or internet fiction, the paper employed a case study of Tomato Novel, a leading platform owned by ByteDance. It explored how Tomato Novel’s advertising-supported free content model, which pays authors based on ‘effective reading time’ spent on their works, is reshaping both publishing economics and storytelling practices. This model incentivises fast-paced, emotionally engaging storytelling designed to maximise reader retention and advertising exposure. At the same time, the platform’s extensive integration of AI throughout the content production process is reducing costs and expanding content supply. While these developments align with the economic logics of platform capitalism and free-content business models, they also raise concerns about market saturation and creative autonomy. The paper argued that Tomato Novel represents both an alternative publishing economy that broadens literary participation and a mechanism through which platform power is consolidated across the creation, circulation, adaptation, and monetisation of webnovels.

You can read more about his project here.

Exploring AI Enabled Author Identities: A Case Study

Dr Frania Hall, University of the Arts London, College of Communication, UK

In her presentation Frania explored the role of the author and the shift in author identity of AI-written books. The presentation focused on a study of Barrett Williams, a systems engineer who has produced more than 4,400 AI-assisted nonfiction titles through a one-person “book factory.” His workflow uses specialized AI agents for topic selection, drafting, editing, metadata, and production.

Amazon limits how many books can be published each week (around 10) which limits how many books Williams can publish each month. The mass production of books has an impact on Williams self perception – he sees himself as a book manufacturer rather than an author and he is very clear that the books have been written by ChatGPT. Frania highlighted that the emerging publishing model is a probabilistic system and controlled by Amazon as the distribution platform.

The system Williams developed, models the traditional publishing model and the author identity that emerges sounds very familiar to any publishing house. He has created a workflow with AI agents that cover what in traditional publishing can be recognised as value chain. There is one agent that selects the topics, one that proofreads and one that copy edits, because Williams realised expertise needed for each step. This cannot be done with just one big generic AI process. Frania highlighted that the work done by AI agents can be mapped onto traditional publishing systems. What Williams has developed is actually a system of publishing that follows the existing publishing model.

Frania concluded that the emerging author identity positions the “author as engineer”, concurrent to traditional publishing. “He sees it as something very separate he’s doing. He is not aiming to compete. He’s in awe of publishers. The identity is consciously separate and different. He is essentially avoiding ethical issues around where this AI content is coming from. He just says I’m separate from the mainstream of publishing, and he doesn’t want to do anything that’s going to get him into plagiarism or anything like that. He describes himself very much as a separate and transparent player. Readers are in control of what they buy, but that side-steps issues around originality and value. And he is, though, despite the slight tensions in what he’s saying, genuine, engaged, and thoughtful in his approach to alternative publishing models.”

Literary Romances and Professional Boundaries: How Publishers of Literature Construct Author Relations

Anne Mölsä, Tampere University, Finland

Anne presented her ongoing doctoral research on the professional culture of Finnish publishers of literary fiction. She reflected on the connections between author relations and professionalism in publishers’ meaning-making and actions. The research data consists of interviews with editorial staff as well as work situation observations in Finnish publishing houses.

In her presentation Anne showed that the publishing profession is centrally constructed by placing the publisher in relation to the author in various ways through boundary work, and her data shows that there are different ways in which authorship is constructed. Firstly, the relationship is constructed as a sort of non-friendship, so the boundary is drawn between the author and publisher by describing the friend-like aspects as not totally genuine, but rather a tool for professional work. Secondly, the author is constructed as someone in need of emotional support, and here the boundary between author and publisher is drawn by constructing the publisher as rational and supportive, and the author as emotional and needy. Thirdly, the boundary can be drawn outside of the traditional relationship. This happens by constructing the author as someone exceptional, especially talented or famous. The professional relationship consists of opposing poles, where there are elements of intimacy, but also of rationality, distance, and control. The ideal of professionalism requires the publishers to place themselves in the intimate relationship, but also to control it at the same time. Authors, on the other hand, are often constructed as not setting as strict boundaries and craving a deeper connection. This can be understood as not sharing the same professional ideals as the publishers. However, the author is also sometimes constructed as a magical being and here the control aspect seems to lose its meaning, as author and publisher form an intimate whole. Anne concluded that publisher and author can be seen as forming a sort of a double helix. They come to being together and depend on each other. So the elements of intimacy and control keep the double helix both strained and energised at the same time. There is a pull to opposite directions.

Around the World in 200 Titles: Mapping Translation Dynamics of Globally Successful Literary, Genre, and YA-Fiction (1990–2019)

Dr Kamila Augustyn, University of Wrocław, Poland

Kamila presented some findings of a translation mapping research project for which she collected data in 2025. The aim of the study is to examine the linguistic scope of selected titles for translation (or ‘coverage’) and the speed of translation by measuring time lags between the original publication and the translations into different languages.

Kamila and her team selected criteria based on author status and literary awards. They investigated how long on average the translation into a specific language takes. Findings show that well-established publishing markets in Western Europe consistently outperform other language markets in translation speed. A limitation was the language diversity of the originals, as 159  were originally published in English. However, Kamila was able to confirm her hypothesis that translational speed depends mostly on the target language.

In summary, the presentation mapped translation coverage and time lags across 200 globally visible titles published between 1919 and 2019. The corpus included 68 literary fiction, 67 genre fiction, and 65 young-adult titles, with 159 of the selected titles originally written in English. The data shows that Western European markets translate titles the fastest. German ranks fastest, followed by Spanish, Dutch, Italian, and French. Findings also show that translation speed improved over subsequent decades with the most important period for translation being between the 1st and 3rd year after original publication. In general, more recent publications tend to generate a higher volume of translations.

The Graphic in Translation

Matthew Chrislip, University of the Arts London, Central Saint Martins College of Art, UK

Matthew presented his PhD research on graphic design and linguistic translation as related practices.

Translation is often casually invoked as both a method and an outcome of graphic design practices. Like linguistic translation, graphic design is concerned with meaning, its forms, and its transferability, and both practices are tasked with reconfiguring and transmitting content across contexts – and, generally, with maintaining some degree of semantic continuity through this change. But as in a geometric ‘translation’, in graphic design this change can be as much material and technical as it is semantic.

In his research Matthew explores graphic design and linguistic translation as proximate and symmetrical practices in order to test the productive potential of this association, as well as to find the limits of its usefulness. This exploration is itself a kind of translation, entailing the comparison of existing forms, production methods, and theoretical frameworks found across the two domains, and supporting the development of new possible forms, methods, and theories.

This research is based primarily in studio practice, where methods of formal production can be deployed as research methods, foregrounding the role of the graphic in the construction of knowledge. The comparison between graphic design and translation is tested through reflexive experimentation with graphic media -especially those concerned with textual material, like typography, book design, and writing practice. These experimental methods are often drawn from specific translation practices: from simultaneous interpretation, an improvisational and open-ended approach to production, and from linguistic translation, an interest in syntax that finds its visual corollary in gestalt principles. This comparison across domains of practice, like all translations, induces a potentially infinite multiplication of semantic and material possibility. But this work finds useful limits through the design of structural frameworks – often books, like the one containing the dissertation of this research – that gather its many trajectories and translate its theoretical propositions into concrete (though provisional) forms.

His projects includs a variable digital font based on the Leroy controlled lettering system, a five-by-eight-inch book titled Simultaneous Interpretation, and an interactive essay capable of nearly 19 million translation permutations. He argued that translation produces multiplicity and requires publishing infrastructures that support improvisation, versioning, and open-ended forms.

You can find out more about Matthew’s project here.

Programming Print Making: Expanding Access and Aays of Working with Print

Oswin Tickler, University of the Arts London, College of Communication, UK

Oswin shared his ongoing exploration into combining creative code and print with the current emphasis on expanding access and methods for printmaking. The presentation focused on the following themes:

1. Workshops
   • Received funding to trial multiple workshops across the UK, focusing on open-source coding languages to make printmaking more accessible.
   • Workshops included combinations of code with digital print, letterpress, and risograph techniques (in addition to his existing workshops in screenprint).
2. Feedback
   • Participants reported increased confidence in coding and were encouraged to think and work differently.
   • Workshop structure was effective, with requests for more advanced levels and continued learning opportunities.
   • Collaborative and experimental elements were highly valued.
3. Future Directions
   • Looking to collaborate with print studios, universities, and arts organizations to expand workshop provision.
   • Some workshop feedback suggested there may also be interest from the secondary education sector due to the way workshops combine art/design/maths/computing.

You can find out more about Oswin’s work here.

AI integration in the European Publishing Industry: A Cross-National Comparative Analysis

Caterina Morelli, University of Rome Tor Vergata, Italy

Caterina presented a cross national comparative analysis that forms part of her PhD research where she mapped AI adoption across the 27 EU member states. Her methodology involved scoping reviews via databases like Scopus and analysing grey literature from national publisher associations between 2023 and 2026. Here AI adoption is understood as a broad analytical construct, an umbrella term for AI use, AI perception, impact, integration, and implementation across the overall publishing industry. Caterina used the European regulatory system (the AI Act) to define what an AI system is which limits the scope of the project to exclude systems in use before generative AI (e.g. for editorial tasks).

The research focused on answering four main research questions:

1. Has the adoption of AI across the EU member state in the broader publishing industry been mapped in scholarly literature?
2. How and to what extent do Publishing Associations across the EU member states documented AI related developments in their sector?
3. How do EU Publishers Associations frame the question of AI?
4. Which theoretical framework could be adopted to interpret differences in how the same phenomenon is framed?

Findings included:

• A sectoral divide where academic and particularly STM (Science, Technology, Medicine) publishers and educational publishers focus on AI-enabled innovation and research tools, while trade publishers focus on operational efficiency and copyright concerns.
• Academic publisher are more keen to take into account the research perspective to create AI-enabled research tools.
• Disparities in accessibility exist: While Italy has consolidated traditions in digital accessibility, other major markets like France and Germany do not yet prioritise AI for accessibility in their surveys.
• AI can be understood as part of society, as part of institutions that make rules, and as technical rules that regulate the interaction between artificial agents similar to the way social institution in traditional social technical systems try to regulate technology and human agents.

Personalizing the Read: Synthetic Voices and the Transformation of Audiobooks§§

Sarah Hedman-Dybeck, Uppsala University, Sweden

As part of her PhD research Sarah examines the use of AI-generated voices in audiobooks, with a particular focus on how these technologies affect perception, the relationship between text and voice, and readers and listeners. In her talk, Sarah presented findings from a survey of 896 Swedish Storytel users who had listened to AI-narrated audiobooks through Storytel’s Voice Switcher.

Most survey respondents valued voice selection and were positive toward AI narration, especially when the voice quality was strong or resembled a favoured human narrator. AI-written and AI-translated literature received more cautious responses, raising questions about whether listeners distinguish narration from the creative and textual work itself. “The voice of the performing narrator is a crucial aspect of the audiobook medium as such, and who is doing the reading and how this reading is carried out naturally affects both the end result and how the audiobook is experienced by reader.” The findings also challenge the assumption that AI enthusiasm is primarily driven by younger users, as the women who are most positive about synthetic voices are in the age group of 50 years and older.

Sarah argued that AI narration will lead to an increase in audiobook with personalisation options being seen as the biggest advantage of the new technology. Storytel’s Voice Switcher was first introduced in 2024, and the feature provides the users with five AI-generated voices to choose from, with four of the voices having a generic but still stereotypical humanised AI profile. Added to this was a voice clone, an authorised voice replica, of Swedish actor Stefan Sauk, one of the most popular performing audiobook narrators in Sweden. It seems that most users appreciate the possibility to choose a voice in accordance with their own preferences. However, the particular bias in the survey data needs to be taken into account, as users who have listened to AI narration are more likely to have a positive response to AI narration than what would be found in a representative sample of the Swedish population. Moreover, Sarah found a higher acceptance of the Stefan Sauk voice clone compared to those who appreciate other AI voices which points towards potential differences in attitudes between voice clones and so-called born AI narrators.

Sarah concluded that audiobook readers have less problem with automating the oral part, the audiobook narration, than the textural parts or the story, and those AI narrated audiobooks appear to be less threatening to human creativity than books co-written or translated by means of AI. This in turn implies that the narration for most reader is not tied to the audiobook as a work of art or an aesthetic object. Most audiobook readers seem to be pragmatics. They like to choose the voice and tailor their own listening experience.