A COMPARATIVE STUDY OF AI-BASED AND HUMAN TRANSLATION QUALITY IN UZBEK–ENGLISH TRANSLATION

Abstract

Recent advances in artificial intelligence have significantly transformed the field of translation, leading to the widespread use of AI-based machine translation tools. While numerous studies have examined the effectiveness of such tools for high-resource language pairs, limited attention has been given to low-resource languages such as Uzbek. This study aims to provide a comparative analysis of AI-based translation tools and human translators in the context of Uzbek-English translation. The research evaluates translation quality across four domains-technical, legal, academic, and literary-using expert-based assessment criteria, including accuracy, fluency, contextual adequacy, and cultural appropriateness. A dataset of selected Uzbek texts was translated using AI tools and by a professional human translator, and the results were systematically analyzed using a Likert-scale evaluation. The findings reveal that while AI-based translation demonstrates acceptable performance in technical and general academic texts, it exhibits notable limitations in handling idiomatic expressions, cultural references, and stylistic nuances, particularly in literary and legal domains. The study highlights the continued importance of human translators for context-sensitive and culturally embedded content and suggests a hybrid translation approach as a practical solution for Uzbek–English translation.