AI-assisted translation of cultural-emotional meaning in Spanish, Uzbek, and English: issues of equivalence and human mediation

Abstract

This article examines the issues of equivalence, pragmatic appropriateness, ethical responsibility, and human mediation that arise in the AI-assisted translation of cultural-emotional meaning across Spanish, Uzbek, and English. The study focuses on selected Uzbek lexical units such as andisha, ko‘ngil olmoq, oriyat, nomus, hayo, mehr, duo, and qarg‘ish; Spanish units including vergüenza, cariño, respeto, alma, corazón, honra, and cortesía; and English items such as dignity, shame, care, respect, blessing, and emotional support. Comparative-semantic and linguocultural analysis demonstrates that, although contemporary AI systems are capable of producing grammatically fluent output, they do not consistently preserve cultural connotations, emotional evaluation, context-sensitive pragmatic meaning, and ethical nuance to the same extent. On this basis, the article proposes an “AI draft – human cultural mediation – ethical validation – final translation” model. The model conceptualizes the translator’s role as extending beyond conventional post-editing, positioning the translator instead as an intercultural mediator and the final decision-maker in matters of semantic and ethical adequacy.