LINGUOPRAGMATIC CONSTRAINTS OF ARTIFICIAL INTELLIGENCE TRANSLATION AND STRATEGIES FOR THEIR MITIGATION
Abstract
Artificial intelligence (AI) has significantly transformed translation practices by enabling fast and large-scale multilingual communication. However, despite its efficiency, AI-based translation systems still face serious linguopragmatic limitations. These limitations are particularly evident in handling context, cultural nuances, idiomatic expressions, and speaker intentions. This paper explores the key linguopragmatic constraints of AI translation systems, including issues related to discourse coherence, pragmatics, and sociolinguistic variation. It also proposes practical strategies to mitigate these limitations, such as hybrid human-AI collaboration, contextual modeling, and improved training data design. The study is based on recent research in computational linguistics and translation studies, aiming to provide a comprehensive academic analysis of the problem.