From the subjunctive to the ergative: A narrative review of AI mediation in Spanish-Hindi literary translation and its impact on translation competence in university teaching

Abstract

Introduction: The expansion of artificial intelligence has transformed translation and translator training, especially through neural networks and large-scale language models. In literary translation, these advances pose additional challenges due to the need to preserve linguistic, pragmatic, cultural, and stylistic dimensions. The aim was to critically analyze the mediation of AI in literary translation, with a special focus on Spanish–Hindi and its educational implications. Methodology: A narrative, qualitative, interdisciplinary, and diachronic review was conducted for the period 1990–2026. The final corpus consisted of 27 studies selected through purposive and theoretical sampling, organized around five axes: linguistic divergence, pragmatic-cultural transfer, technological evolution, translator competence and cognitive load, and voice, authorship, and agency. Results: The evidence showed that technological advances do not eliminate linguistic divergences nor guarantee stylistic preservation. LLMs expand creative possibilities, but they can also introduce unjustified expansions. Furthermore, technological development shifts human intervention from correction to evaluation, verification, and decision-making. In education, this process demands strengthening critical AI literacy, metacognitive regulation, and argumentative skills. Conclusions: Spanish-Hindi literary translation remains insufficiently studied. Specialized corpora, empirical evaluations, and educational designs that conceive of AI as a heuristic mediator under human governance are needed.