TRANSLINGUA https://ojs.aiou.edu.pk/index.php/tra <p>Journal of Translingua</p> en-US Fri, 30 May 2025 00:00:00 +0000 OJS 3.3.0.21 http://blogs.law.harvard.edu/tech/rss 60 Evaluating the performance of ChatGPT, Claude, and Grok on Urdu Text Generation and Language Processing: A Comparative Analysis https://ojs.aiou.edu.pk/index.php/tra/article/view/3215 <p>The study intends to carry out an extensive contrastive analysis of three leading AI language models, including ChatGPT, Claude, and Grok, in their ability to generate, process, and improve Urdu texts. The importance of AI tools in the recent era is of great concern especially with reference to natural language processing. The current study aims to analyze the performance of these AI tools on Urdu language processing since Urdu language has a different sentence structure and vast grammatical rules which are quite different from English language.&nbsp; The study will be conducted in a systematic manner where each tool will be evaluated on its performance with respect to spelling correction, identification and rectification of grammar errors, generation of synonyms and antonyms, paraphrasing abilities, sentence restructuring skills and context specific cultural sensitivity while processing Urdu texts. A corpus will be compiled comprising of Urdu Literature, newspaper articles, academic content, social media content and poetry. A mixed-method approach will be used to evaluate the performance. The assessment of accuracy will be quantitatively measured and the analysis of linguistic behavior will be assessed by qualitative methods. AntConc will be used as a tool to check the concordances for context based cultural sensitivities. The results will be part of the knowledge of strengths and weaknesses of each platform, providing practical recommendations to educators, researchers, content developers, and language technology developers dealing with Urdu text. The results of the research will provide recommendations on the best practices of applying these AI tools and also reveal the areas that still need more work in the Urdu natural language processing.</p> Fariha Saghir, Maria Sundas Copyright (c) 2025 TRANSLINGUA https://ojs.aiou.edu.pk/index.php/tra/article/view/3215 Fri, 30 May 2025 00:00:00 +0000