In 2023, UK consumers spent £26m ($33m) on translated fiction, marking a 12% increase from the previous year, according to Nielsen.
Foreign fiction in English-speaking countries has been steadily gaining popularity, with 13 of the 100 novels on The New York Times’ ‘The 100 Best Books of the 21st Century’ list being works of translated fiction. Elena Ferrante’s My Brilliant Friend (originally L’amica geniale in Italian) was featured as the number one book on the list, highlighting the growing demand and appreciation for translated fiction in literature.
Translated fiction allows readers to engage with texts from other cultures and provides authors with the opportunity to expand their audiences. Different countries have distinct literary styles, and for readers who do not understand the original language, translated fiction makes these novels accessible.
Yet, the work of the translator can be overlooked, and this is being exacerbated by improvements in machine translation. For example, a Society of Authors survey published in 2024 found that over a third of translators have lost work due to generative AI. While technology can assist in breaking language barriers, it often lacks the nuanced understanding and cultural sensitivity that human translators bring to the table.
Experimenting with AI-translated fiction
In November 2024, it was announced that the Dutch publishing house Veen Bosch & Keuning (VBK) would be experimenting with the use of artificial intelligence to translate a small number of novels. Notably, this initiative focuses on commercial fiction, such as crime thrillers, romances, science fiction, and fantasy, rather than literary fiction.
This decision was met with outrage by translators and authors alike. Not only does this downplay the intricate work done by translators, but it also seems to undervalue commercial fiction.
Michele Hutchison, the winner of the 2020 International Booker Prize for her translation of The Discomfort of Evening from the original Dutch, expressed concerns that this implies such books “are purely formulaic and don’t contain many creative elements”, which is an insult to both authors and readers of these genres. Furthermore, as the main benefits of AI translations are time and cost savings, this suggests that these novels are not worth investing in.
While AI is an improvement on previous machine translators, it still has its flaws. Critics of AI translation argue that the output is bland, unidiomatic, and lacking in emotion. AI struggles to recognise tone, meaning important qualities like sarcasm and humor can be lost in translation. One proposed solution is to have AI perform the initial translation, which would then be refined by a human.
However, this approach may not save much time in the long run, especially if the AI’s output requires significant revision. Studies have found that this approach can actually take much longer, as fixing something done badly but quickly is often more time-consuming than just doing it properly in the first place. In that case, why not just have the human translator handle the entire process from the start?
AI and accessibility
The backlash VBK’s decision received suggests that there may not be a market for AI-translated fiction. Consumers can find AI-generated literature unappealing, and the same sentiment extends to translations. There is an argument that AI translation tools could enhance the accessibility and popularity of novels written in minority languages, particularly those that lack human translators.
However, many AI models are primarily trained on English language data, with a relatively small fraction of the total dataset consisting of foreign language data. As a result, English is essentially AI’s first language. Subsequently, AI may not be able to translate minority languages, which it has far less experience with, to an acceptable level, instead generating gibberish and misrepresenting the author’s true intent. Rather than making these novels more popular, then, AI translations could have the opposite effect—misleading the readers into thinking that the book itself is bad, rather than a bad translation of it.
Furthermore, AI-generated translations could depreciate translation work. Machine translators and computer-assisted translation tools have already become widespread in commercial translation, leading to a decrease in available work for human translators. Ian Giles, co-chair of the Translators Association, warns that AI will exacerbate this trend, resulting in a situation where “only those with wealth [will be] able to translate literature for publication.” Therefore, while using AI as a translation tool may enhance accessibility for certain novels, it could simultaneously render the translation industry itself far less accessible.
The human element
Translation is far from a mechanical process; it’s not as simple as looking up words in a dictionary. Elements like syntax, tone, rhythm, and wordplay all play a crucial role in shaping the overall effect of a piece of literature but do not always translate well into other languages.
Translators do more than just convert words; they translate feelings and emotions. They must produce a consistent voice that complements the original work while making idiomatic sense in the target language. The process is complex, and translations can be considered works of art in their own right.
Ultimately, AI may appear to speed up translating, but in the process, it removes the human element—the very reason these works of fiction are so highly valued.
In an open letter to VBK, the European Council of Literary Translators’ Association stated, “Books are written by human authors and should be translated by human translators. Imagination, understanding, and creativity are intrinsically human and should not be left out of any literary text.”
The human touch in translation ensures that the subtleties and emotional depth of the original text are preserved, and if we want to ensure that global fiction continues to fill our bookshelves, we cannot relegate the task of translating it to a machine.
