Fewer than 10 titles will be translated, The Guardian reported, all commercial fiction — as opposed to literary titles — and only books whose English rights have not been sold. “We don’t foresee the opportunity to sell English rights of these books in the future,” VBK Commercial Director Vanessa Van Hofwegen was quoted as stating.
“Have you always dreamed of being read in English?” a VBK notice asked Dutch authors. It went on to explain that the English translations will be made available as e-books, with the first titles to be released in the summer of 2025.
Slator Translation as a Feature (TaaF) Report
The Slator Translation as a Feature (TaaF) Report is a vital and concise guide on how AI translation is becoming an integral feature in enterprise technology.
Ian Giles, Chair of the Translators Association at the Society of Authors, called this “concerning news,” citing a SoA survey published earlier in 2024 that found “one-third of literary translators are already losing work to AI.”
“Where work itself is not lost, translators struggle to increase their prices in the face of the AI challenger,” Giles told The Bookseller. “If this publisher feels the need to consult human translators or editors to adjust the output, they are recognising the flaws in this approach.”
Commenter Jane Davis agreed with Giles: “On the few occasions I’ve agreed to do post-editing, I’ve been paid maybe a quarter of my translation rate.”
“Simon & Schuster doubtless think this will save them money?” commenter Louise Rogers Lalaurie replied. “Simon & Schuster should also have estimated the cost to them of the reputational damage their disastrous decision will surely wreak.”
On X, Dutch speaker Sarah Thurley shared, “I pity the poor English-speakers who will be palmed off with mechanical AI translations. Not to mention the translators who literally have to rewrite but are paid peanuts for ‘revision’.”
Happy End-Clients (a.k.a. Readers)?
Dr. B.J. Woodstein, a professor, writer, and Swedish-English literary translator, explained in an April 2024 SlatorPod episode that the SoA establishes a recommended minimum rate for professional literary translation.
“Of course, you meet people who are so desperate for work or really want to work on a particular project that they say, ‘Yeah, I’ll do it at a cheaper rate.’ […] But then you have to think about what the quality’s going to be there,” she said.
The same concerns about quality were borne out in an October 2024 paper from the Natural Language Learning and Generation Lab and the University of Aberdeen, which concluded that literary translation remains “an exclusive domain of human translators.”
In the study, four student annotators and four professional translators expressed a clear preference for human translations over even the best-performing large language models (LLMs). (GPT-4o was ranked a close second, described as “closely approaching human literary translation,” followed by Google Translate and then DeepL.)
Evaluation metrics for machine translation (MT) present a challenge in evaluating the quality of literary translation by LLMs. Standard MT evaluation metrics operate at the sentence level, without taking into account greater context. Newer LLM-based metrics, meanwhile, remain largely untested on literary texts.
Until better evaluation metrics are established, the authors warned, overestimating the quality of LLM-enabled literary translation could lead to the displacement of human translators, reduced salaries, and, ultimately, a decline in the quality of translated works.
Woodstein told Slator that she and other literary translators have already begun receiving queries from publishers to edit and clean up machine-translated texts — for a much lower fee than usual.
“Translation is an expensive cost, we can’t deny that, but in their eagerness to [cut costs], they are not thinking about that end-quality,” Woodstein said. “From a practical perspective, it’s like, ‘Are we going to have jobs in a few years?’ Because AI is going to get better.”
