Semantic Change And Drift In The Urdu Language: A Critical Analysis Of Ai Translation And Large Language Models

Abstract

The study delves into Semantic Drift and A Critical Analysis of AI Translation and Large Language Models. The change in the Urdu language with special attention to how the meanings of some common words have evolved and how Artificial Intelligence large language models like the most widely used translating engine, Google Translator, fail to translate culturally appropriate meanings. Ten words were taken for analysis in this research paper to show different significant changes in their semantics from their initial use in Urdu, comprising "لوٹا" (Lota), "خواب" (Khawab), "غیرت" (Ghairat), "شہرت" (Shohrat), "حسن" (Husn), "بےحس" (Behas), "جاہل" (Jahil), "آبرو" (Aabroo), "شعور" (Shaoor), and "عزت" (Izzat). The study demonstrates how Urdu changes across time concerning [1]its environment, i.e. society, culture and also new words coming in. From the semantic point of view, these changes demonstrate that language is not just subject to changing conventions; it also expresses wider cultural & social transformations, including issues related with politics on both a national and international level. In addition, this research evaluates the weaknesses inherent within artificial intelligence-oriented approaches towards cross-cultural meaning interpretation and translation thereof. By so doing, we can see that there is still much which is left unknown about the development of languages and there should be some improvements done in this area in Large Language Models by feeding cultural interpretations into their deep database, taking into account all the peculiarities of different cultures and languages.