A new 150-page research report, Books in Translation: Trends and Transformations in the European Publishing Market, finds that translations from English account for 50% to 70% of all translated titles across most of the continent. It also finds that translations from Japanese are proving more popular than books from some European languages, and that AI is already having a significant impact on the translation market.
The report draws on data from 10 countries and regions, combining national library statistics from Sweden, Slovenia, Catalonia, Bulgaria, and Croatia with publishing industry data from Germany, France, Italy, Poland, Spain, and Austria. It was produced by Rüdiger Wischenbart Content and Consulting in cooperation with CulturalTransfers.org and is part of the EU-funded ThinkPub project.
Translations continue to play a significantly larger role in European book markets—especially those with fewer than 10 million speakers—where they typically account for 20% to 40% of annual book production. In larger markets such as Germany, France, Italy, and Spain, local-language titles dominate, and translations represent roughly 12% to 15% of new books. Direct translations between smaller European languages, such as Slovenian and Bulgarian, remain “almost non-existent” and are often derived from an English translation first, with English serving as a gateway language.
One notable shift in the translation market has been the growing popularity of Japanese manga. In France, Japanese is now the second most translated language, representing 19.9% of all translated titles in 2024. In Germany, manga overtook French in 2021 and accounted for 12.8% of new translated titles in 2023. Poland recorded 11% of translations from Japanese, also exceeding French.
All of this is unfolding within an increasingly difficult publishing environment. The report notes that annual unit sales are declining across almost all markets over the past two decades, with the notable exceptions of Portugal and Spain. Over the past two decades, European publishing has lost 24.1% of its real market value, once revenue is adjusted for inflation, according to the Federation of European Publishers.
On the policy side, the report finds that EU translation grant funding through the Circulation of European Literary Works Programme has increased from approximately €3.4 million annually in 2016–2017 to as much as €5 million per year from 2021 to 2024. It also notes that translations from Ukraine were the fastest-growing category, with the EU supporting the publication of 130 books—representing 6% of all funded works between 2021 and 2024. This boom came in response to Russia’s invasion of Ukraine in 2022.
The report also highlights sharp imbalances between national programs. It cites Slovenia (population 2.1 million), which publishes 3,800 original works annually and supported the translation of 165 titles from 2020 to 2024. By contrast, Bulgaria (population 6.3 million), which produces nearly twice as many books each year, supported just 18 titles through its national translation funding program.
AI's impact
The report also examines the impact of AI on the translation market, identifying two events as pivotal. First, in November 2025, Amazon introduced Kindle Translate, an AI-powered translation tool available exclusively to self-publishing authors on Kindle Direct Publishing, covering English-to-Spanish, Spanish-to-English, and German-to-English language pairs. The report suggests that the tool’s current restriction to larger language pairs—reflecting limitations in training data—is likely to widen the gap between dominant and more peripheral languages.
A second development, noted by the report, came in February, when HarperCollins France partnered with Fluent Planet to use AI to translate the Harlequin Azur romance series. HC justified the move by arguing that, in light of declining sales, it was necessary to preserve the €4.99 price point. Predictably, the decision drew pushback from French professional translation bodies and reignited debate over AI’s role in literary production.
To test the practical limits of machine translation, the report’s authors conducted a small experiment, translating a Slovenian text into Zulu via ChatGPT-4 in Ljubljana and then back into Slovenian and English via a ChatGPT account in Pretoria. The locations were chosen because there are no direct Slovenian–Zulu human translations, dictionaries, or documented cultural contact between the two languages. The AI performed reasonably well on short, technical sentences but struggled with complex literary passages, omitting details and altering meaning.
A subsequent test on a Slovenian poem found that while the direct Slovenian-to-English translation was adequate, the Slovenian–Zulu–English pathway produced what the authors describe as “a new poem”—one wholly transformed and stripped of its cultural and political subtext. The authors concluded that AI translation is likely to replace a number of translators, but will still require highly skilled human editors to catch errors, especially in literature and in translations involving languages with limited training data.
The report’s authors are direct about what this shift means for the profession. “In the era of machine intelligence,” they write, “only human authors and translators with above-average skills will be able to produce new content; the average ones will be substituted by the machines.” They frame this not as a warning but as a consequence—one that demands more rigorous training, higher editorial standards, and a clearer-eyed publishing industry. “And this,” they add, “is not bad at all.”
Authors of the report are Miha Kovač, professor at the Department of Library and Information Science and Book Studies at the University of Ljubljana; Austrian publishing consultant Rüdiger Wischenbart; Yana Genova, founder of the Next Page Foundation; and Anja Kamenarič, researcher at the University of Ljubljana.
