In February 2022, while working on translating American writer Dana Spiotta’s novel Wayward into French, literary translator Yoann Gentric decided he needed a break. He wanted to test whether AI could replace his job.
Gentric had been struggling with a short, wordless sentence that described the main character’s feelings when she opened a window: “Bright, sharp night air, bracing.” He entered this into DeepL, a machine translation tool powered by neural networks that often beats Google Translate in accuracy tests.
The translation it gave was reassuring for his job security: L’air de la nuit, vif et vif, était vivifiant (The night air, lively and lively, was enlivening). The AI captured the meaning but seemed unaware that the repetition made the line sound ridiculous. It was far worse than his own translation, which would appear in the book a year later: L’air pur et piquant de la nuit, vivifiant.
When Gentric repeated the experiment this spring, the result made him less comfortable. This time, DeepL suggested: L’air nocturne était vif, pur et vivifiant. The online translator still lost the sentence’s style by adding a verb, but it had learned to use three different words that even had a musical quality. “I don’t know if it’s just luck or a fine-tuned algorithm at work, but nocturne and pur isn’t bad,” Gentric said.
Chatbots powered by large language models (LLMs)—neural networks trained on huge amounts of text to produce natural-sounding language—are quickly becoming part of every aspect of our work and free time. But few professional fields are being disrupted as fast as the translation industry in Europe, which has over 200 languages and a thriving tech sector.
According to a recent joint survey by French authors’ societies ADAGP and the Société des Gens de Lettres, 79% of translators believe the rise of AI “poses a threat of replacing all or part of their work.” In Britain, a 2025 survey found that 84% of translators expected lower demand for human translation, leading to lower pay.
These fears are about the future, but for many translators, their work has already changed. Laura Radosh, a German-to-English translator based in Berlin, used to get about four job offers a month from clients like universities, professors, and museums. Last year, that dropped to one per month.
Many of those jobs were “post-editing,” which meant correcting texts that had already been run through a machine translation engine. “Post-editing took me as much time as translating from scratch,” Radosh said.
Post-editing is far less creatively satisfying than translating from scratch, and it also pays less. It’s usually paid by the hour rather than by the page or the book, and according to the French translators’ association, it’s paid “at unacceptable rates considering the work involved.” In Germany, publishers have been known to offer typical rates of two to eight euros per page—a quarter of the average pay for translating a page from scratch.
But rates for regular technical translations have also dropped. “I was offered a job at 60 cents a line,” Radosh said. “Before that, 80 cents was the lowest rate I had ever seen.”
Even before LLMs came along, translation was an unstable profession. A recent survey by the German translators’ association VdÜ found that the average income for literary translators—traditionally among the lower-paid in the field—was as low as €20,363 per year before tax. But the latest changes in the industry mean that for many translators, the numbers no longer add up. Radosh recently took a part-time job doing bookkeeping.For an NGO.
Europe’s AI translation industry has been warned that partnering with US companies could damage its reputation.
Read more
Marco Trombetti, co-founder and CEO of the machine translation company Translated, said: “Without help, the human brain can produce about 3,000 words of translation per day. Beginners manage around 1,500, and the best translator in the world might reach 6,000. The difference isn’t that big.”
He argued that the cost of human translation has so far been determined by the number of neurons in our brain. “That’s about 100 billion,” Trombetti said. “But if we change that, we change the basic economics of translation.”
However, the rapid pace of technological change is also highlighting what human translators still do best. For one thing, many machine translators still struggle with context. The German-British academic publisher Springer Nature offers its authors the option to have their books automatically translated into other languages for free. But despite promises of later “human checks,” this process has sometimes led to funny mistakes.
In 2024, Springer Nature used machine translation to turn an English book by a group of Indian academics called ‘Capital’ in the East: Reflections on Marx into German. In the chapter headings, the machine translator DeepL translated “capital” not as Kapital (meaning capital in the economic sense), but as Hauptstadt, which means “capital city.”
A spokesperson for Springer Nature said in a statement: “Our AI-supported translation is led by humans and reviewed by professional editors. Errors like this are rare and regrettable, and this case was part of a limited pilot that has since ended.”
Jörn Cambreleng, director of Atlas, a French organization that promotes literary translation, said: “Machine translation is not creative. These systems are built to produce generic sentences—sentences that have been said before or sound like they have. Good human translators, on the other hand, try to put into words something that has never been said before.”
View image in fullscreen
Katy Derbyshire: “I understand what someone might scream when they hit their toe on the bed frame—an algorithm doesn’t.” Photograph: Nane Diehl
One of the ironies of this upheaval is that literary translation now seems like a safer career choice than technical translation.
The HarperCollins-owned imprint Harlequin France has confirmed that it is working with a French communications agency, Fluent Planet, to produce translations generated by AI software and then edited by humans. For now, though, these trials are limited to the less literary end of the market: Harlequin’s titles include A Mistress’s Confession and The Embrace of a Prince.
In Germany, where the total number of new published books has been slowly declining each year, translated literature has held up remarkably well. In 2024, 8,765 translated books were published, making up a historically high 15% of all new books. Increasingly, authors are also contractually requiring their publishers not to use AI in the translation process, said Marieke Heimburger, a Danish-to-German translator who chairs VdÜ.
“AI really can’t handle dialogue,” said Katy Derbyshire, a Berlin-based translator who has turned novels by Clemens Meyer, Christa Wolf, and others into English. “When you translate from scratch, you learn to understand the characters and their motivations, and you’re constantly adjusting them in your head—to specific situations and to genre. The dialogue AI came up with just didn’t fit the character description at all.”
Being human helps the translation process, she added. “My body has experienced all the pain and joy that literature tries to convey. I understand what someone might scream when they hit their toe on the bed frame—an algorithm doesn’t.”
Fernando Prieto Ramos, from the University…The head of the University of Geneva’s faculty of translation and interpreting said his center noticed a drop in applications to translation courses three years ago, when the rise of generative AI fueled excitement around machine translation. “But the trend is gradually reversing again, thanks to a more diverse range of training options,” he said.
Even those who develop machine translation software admit there are tasks it still can’t handle. “If in Italian I say ‘Solo tre parole: non sei solo,’ a literal translation into English would be ‘Just three words: you are not alone,'” said Trombetti, who founded Translated in 1999. “But that gives you four words, not three. That’s something machine translation still struggles with.”
Heimburger said: “I’m not really afraid of AI, because I know it can’t do what I can do. What I’m afraid of are the people who think AI can do my job.”
Frequently Asked Questions
Here is a list of FAQs about the article Being human helps despite the rise of AI is there still hope for Europes translators
BeginnerLevel Questions
1 What is the main point of the article
The article argues that while AI is changing the translation industry human translators still have a futureespecially for complex creative or sensitive work The key advantage is being human
2 Is AI going to replace all human translators
Probably not AI is great for simple repetitive translations But for nuanced texts like literature legal documents or marketing humans are still needed for accuracy tone and cultural understanding
3 What does being human helps mean in this context
It means that humans can understand context emotion humor and cultural references in a way AI currently cannot For example a human knows when a joke is appropriate or when a phrase has a hidden meaning
4 Is it still worth studying to become a translator
Yes but with a twist Youll need to specialize in areas where human judgment is critical and learn to use AI as a tool not see it as a threat
5 What kind of translation work is safest from AI
Creative translation legal or medical translation and any work requiring deep cultural knowledge or emotional nuance
IntermediateLevel Questions
6 How is AI actually used by translators today
Most translators use AI tools to get a rough first draft Then they edit and polish the text adding context fixing errors and adjusting tone This is called postediting
7 What are the biggest weaknesses of AI translation
AI struggles with idioms sarcasm wordplay and texts with multiple meanings It also fails in highly regulated fields where precision is everything
8 Can AI handle regional dialects or minority languages
Poorly AI is trained on large datasets so it works best for major languages For smaller languages or dialects human translators are still essential
9 What does postediting mean and is it a good job
Postediting means taking a machine translation and fixing it
