“Love grows stronger when its form seems modest, and devotion remains true even if it wears a quiet face,” one voice insists.
“Love may appear fragile yet holds steady; it does not diminish when it is shown less. For some, that affection is a financial transaction—visible, outspoken, and proclaimed by language,” another voice adds.
And a third: “Love may look weak, yet its strength remains, and it is not love that changes because appearances lean one way.” the author notes.
It seems clear which version was written by a man and which by a machine. Artificial intelligence enters the scene as a translator, a field notorious for pushing people to question what they can or cannot translate. In a moment when many articles celebrate or fear AI, the tricky craft of translation stands out as a perfect testing ground—capable of cutting costs and easing workflow, yet prone to dissatisfaction if quality dips.
Jorge Carrión, co-author with Taller Estampa and GPT, speaks of artificial writing theories and translation programs as “useful tools for quick drafts of screenplays, technical works, or essays without literary ambition.” Human oversight remains indispensable. Machines may draft, but they cannot yet capture the full depth of high literature, though that gap could shrink in time.
Pia Gepetto—a name given to ChatGPT for this report—offers a cautious view: “Machine translation tools have progressed and can produce accurate and readable literary translations in many cases. Literary translation, however, remains a highly specialized field that requires a profound sense of language, including nuances, puns, double meanings, and metaphors.”
Ernest Folch, editor of Folch & Folch, expresses skepticism: translating heavily technical or mechanical texts with AI is only vaguely possible. A translation, like any creative work, needs emotion, sensitivity, and good judgment that a machine cannot possess. The human touch is essential, especially in literature where tone, rhythm, and cultural nuance matter.
Ernest Folch, editor: “A translation, like any creative work, needs emotion, sensitivity, and good judgment, all the virtues a machine can never have.”
Cupcake Palmer, winner of the XXIV Ángel Crespo Translation Award, views artificial intelligence as a practical aid for technical work but warns that applying it to literature is a mistake. Even when machines assist, expert revision is necessary, and only seasoned human eyes can detect and correct serious errors. Literature involves choices about sentence rhythm, metaphor, pun, dialogue, and cultural bias that machines struggle to emulate.
reduce costs
Yet some publishers press for translators to act as proofreaders for AI-generated translations. They argue it lowers costs, but the result may be a trade-off between price and quality, not an ideal for readers seeking meaningful literature.
Magdalena Palmer, translator: “Artificial intelligence is a useful tool for technical translations, but applying it to literature is a mistake.”
For translator Cristina Macia, translating literature by machine is “not possible” and has repeatedly ended in disaster. If future breakthroughs occur, imagine a team of four irreplaceable translators for deluxe editions, where every line is still a human decision. Such editions would be marketed as eco-friendly products, yet the real value would rest on human expertise and nuanced interpretation.
victims
Meanwhile, industry chatter reveals a tension: some publishers reportedly translate simpler works with software, relying on machines to handle the basics while human editors polish the output. If true, it mirrors old stories about witches—there are no books translated by AI, yet some are. The first victims would be translators and readers who deserve quality.
Industry gossip speaks softly of a publisher who translates less demanding books using software.
Author Patrick Pron notes that not all publishers respect the profession: a good translator is a strong commercial asset and a guarantee of quality. Some editors are willing to pay translators less, betting on AI to cut costs. For those who must compete with digital workers in precarious conditions, the news feels alarming. Readers could end up with books that lack depth and care.
distinction
Carrión senses a moment when it will become impossible to tell whether a translation came from neural networks or human hands. Gepetto also suggests that as tech advances, the line may blur. The outcome will depend on the work’s complexity, and on how deeply a translator or writer understands cultural and historical context. It may not even be worth asking who translated something if quality is consistent.
Perhaps the great danger lurking in automatic translations is that at some point they will become the standard of quality.
In a recent discussion in a major newspaper, Noam Chomsky remarked that a fundamental mistake of AI lies in treating the human mind as a vast statistical engine. Real understanding comes from a smaller, elegant system, not from data-hungry models that surface likely answers. The insight emphasizes human expertise over raw computation when it comes to creative work.
neoliberal strategy
Publishers may share a similar worry. The fear is that technology worship and cost-cutting rhetoric push people to believe machines can enhance human cognition. The truth is that literature demands emotion, sensitivity, and common sense—qualities a machine never truly possesses. The idea that AI could produce truly literary translations is still a promise with little substance. The strongest gains appear in technical fields, not in artful storytelling.
In any case, a real risk looms: automatic translations becoming the default measure of quality. When that happens, publishers might ask human translators to imitate machine efficiency. There would be no heroic rescue by a Terminator; instead, a quiet drift toward standardization, where the human touch is increasingly rare.
Note: the examples above illustrate translations of a Shakespeare sonnet chosen for this report; the first translation belongs to a renowned author, the second to a modern AI tool, and the third to a contemporary editor. These voices demonstrate the range of outcomes seen when literature meets technology.