A Corpus-Based Sentiment Analysis of Okara University Students' Attitudes Toward Artificial Intelligence Tools in Higher Education

Abstract

This study is a multiple-methods empirical study analyzing the perspectives of University of Okara students on the integration of generative AI tools in higher education, in light of the policy directive by the Higher Education Commission of Pakistan, which mandates students to be AI literate for all public sector universities by the end of 2026. The adopted approach is a mixed methodology consisting of a quantitative survey of 45 postgraduate academics (MPhil) and a qualitative computation corpus-based sentiment analysis of 60 reflective essays written by students. As shown in the quantitative analysis based on the Technology Acceptance Model, 88.9% of the respondents agree that AI technologies improve their conceptual understanding and save time in completing tasks. These technologies are largely accepted for their pragmatic benefits, but fundamental concerns around cognition and ethics restrict the acceptance since 95.6% of researchers believe in the possibility of severe intellectual overdependence and 75.6% more believe that these technologies cause academic dishonesty. The qualitative corpus is analyzed with keywords in context. The positive evaluations in the computational analysis are structurally related to responsible human agency and precise prompt engineering, while the negative evaluations are structurally related to the loss of students’ voices and digital isolation with passive learning. Data indicate that students employ conversational agents as scaffolding devices for operation, but they are worried about cognition offloading and academic integrity. The study concludes with strategic institutional recommendations and clear ethical disclosure standards for balanced and responsible use of technology in Pakistani higher education.