Student perceptions of the impact of AI tools on academic performance: Evidence from Sri Lankan state universities

Abstract

The rapid integration of Artificial Intelligence (AI) in global education has generated considerable interest in its potential to transform learning experiences through personalised pathways and skills forecasting; however, there is a lack of research investigating how students in Sri Lankan state universities perceive AI integration, particularly in resource-constrained contexts. This qualitative study examined undergraduate students’ perceptions, experiences, and expectations regarding AI tools, particularly ChatGPT, Grammarly, and Quillbot, with a focus on personalised learning and skills relevant to future careers. Data were collected via semi-structured Zoom interviews with 20 purposely selected students from three state universities, guided by the Technology Acceptance Model and Self-Regulated Learning Theory. The data were transcribed, anonymised, and evaluated thematically using Lichtman’s 3C framework. Research indicated predominantly positive perceptions of AI as a mechanism for switching from passive, lecture-based instruction to interactive, personalised learning, with features like automated feedback and intelligent guidance appreciated for their potential to tailor material to individual needs. Students had a marked preference for AI tools, which significantly enhanced comprehension of lecture content, aided in literature searches, and provided information for projects; nevertheless, their effect on the completion of writing tasks was relatively limited. Infrastructure constraints have been discovered to be persistent challenges to the effective use of AI, especially at universities situated in rural regions where access to technology, internet connection, and technical help is often restricted. Participants showed significant interest in AI-driven skills forecasting to enhance the alignment of academic education with job requirements, while underscoring the need for culturally contextualised, student-centred AI solutions specifically designed for the Sri Lankan context. The study emphasises the necessity of integrating AIenabled technologies into curricula, along with faculty training and strong ethical governance frameworks that guarantee transparency, in addition to policy initiatives aimed at enhancing digital infrastructure in underdeveloped regions and advocating localised AI solutions to promote inclusive and equitable education..