A letter published by the Association des traducteurs littéraires de France, also known as l’ATLF, and the collective En chair et en os, stated that dozens of Harlequin France translators were informed that their contracts would be ending as soon as possible. HarperCollins, Harlequin’s parent company, confirmed they would be using an AI-assisted translation workflow with Fluent Planet rather than relying primarily on human translators to translate their books. The company pointed to their declining sales as the reason for partnering with Fluent Planet. “Sales of our Harlequin collections have been declining in the French market in recent years,” the statement read. “We want to continue offering readers as many publications as possible at the current low retail price.”
A Fluent Planet spokesperson told The Bookseller that the company uses a hybrid of “in-house language assistance tools with systematic human translation carried out by professional literary translators.” Their website states that the human translators have “studied at the best translation universities or have decades of experience under their belt … When compared to standard machine translation systems that use neural networks, BrIAn [the Fluent Planet AI software] can provide 2 to 3 times higher quality translations, that are more accurate, offer idiomatic phrasing, provide a deeper understanding of the meaning and a faithful representation of the style and emotions of the source text. BrIAn takes into account the author’s tone and intention, making it highly effective for complex literary or marketing content.” Fluent Planet appears to believe their software is on par with human translators while also saying they’re offering a tool not intended to replace human expertise or editorial judgment.
Much can be said about the ethical issue of firing people so they can be replaced by AI. However, I want to focus on the nature of translation and why as of now, even Fluent Planet admits, AI can’t do it well, otherwise they wouldn’t need human translators with decades of experience to check their work. Translation is complicated, delicate work that requires deep knowledge and understanding of both languages. It’s about nuances and subtleties—two key factors that AI is notoriously bad at. When there’s a lack of nuance and subtlety, a great deal can be distorted or lost. Translation isn’t simply about substituting words. It requires the human touch to most truthfully convey idioms, jargon, regional dialects, and even with humans the results are never perfect because that’s the challenges of language. There is never a true one-to-one.
As someone who grew up bilingual, speaking both Spanish and English, I’ve always dealt with the frustration of needing to translate one thought from one language to another. Even when there’s a word or phrase that appears to have a direct translation, it doesn’t often hold the same meaning or intensity I’m trying to express. This is a difficulty I feel in minor translation moments; I can’t imagine needing to translate a whole novel. To translate a novel, the translator must first understand it in its entirety—the themes, the character arcs, the central truth. When there can’t be a 100 percent word-for-word accurate translation of a book, then it becomes about being as faithful to the original work as possible. AI cannot yet do this in the same way humans can because it lacks lived experience and interpretive judgment.
When I read the news about HarperCollins decision, I couldn’t help but think of R.F. Kuang’s “Babel, or The Necessity of Violence: An Arcane History of the Oxford Translators’ Revolution,” a book that utilizes language and translation to explore narrative themes of colonialism and exploitation. In Babel, our main characters contend with the brutality of translation. It shows how language can be weaponized, how the act of translating can be a betrayal of the original, violent and sharp, harming those at the hands of the people in power. Babel examines the harm of one group of people controlling the narrative, deciding which languages are prioritized over others. There is a reason that access to reading and writing has been used as a tool for oppression. Having access to knowledge gives you agency, when that knowledge is taken away, restricted, or watered down it prevents people from understanding themselves and each other.
Language is a constant shifting force, which is why dictionaries are constantly adding new words. It changes with the times, with environment, with society—always adapting. Something that is consistently changing and created by humans, can only be fully understood through the human experience. Even then language is a deeply internal thing. Emotions and thoughts can’t always be perfectly conveyed from one person to another even when they are both fluent in the same language—how can a software ever get to the heart of that? The act of translating one book from one language to another is trying to bridge the gap from one human experience to another and it’s not something that can be reduced to an algorithm. By its very nature it must be different and specific to each book.
