g10
Abstract
Artificial intelligence, particularly through the emergence of neural machine translation systems, has revolutionized language translation. These tools process vast amounts of text at astonishing speeds, delivering a level of accuracy that would have seemed almost science fiction just a few decades ago. However, beneath this technical prowess lies deep limitations, especially when it comes to literary works. Consider, for example, poetry or novels. While AI can certainly extract and translate the essential meanings, many of the subtleties of the original work are lost in the process. Literary translation requires more than simple linguistic equivalence. It relies on sensitivity to the text, an understanding of cultural references, an understanding of tone, and an appreciation of stylistic details—none of which AI can currently truly emulate. Human translators combine these elements to recreate the “art” in a new language, preserving the rhythm, creative metaphors, or unique emotional resonances that define literature. In contrast, current machine translation models tend to flatten language, reducing it to its most literal interpretation. This can lead to an erosion of the subtlety and complexity that scholars and readers have long valued in translated works. The spread of AI translation therefore raises legitimate concerns about the potential loss of the subtlety, precision, and cultural depth that human translators have traditionally preserved in world literature.