AI Driven Personalized Learning and Learning Analytics in Higher Education: Effects on Student Engagement, Achievement and Research Integrity

Abstract

The study aimed to examine the effect of personalized learning through artificial intelligence (AI) and learning analytics on student engagement and learning outcomes of higher education institutes in Pakistan with implications for Research Integrity at these institutes. Adopting a mixed methods approach, quantitative data gathered from 500 students from an undergraduate science program, with semi structured interviews with 30 students for the qualitative data. The descriptive statistics indicated that the mean engagement scores for those students who have access to AI based personalized learnings were significantly higher than the mean scores of students in traditional learning setting; it is from these that an increase in engagement by the students were observed. There was an improvement in Academic achievement as well, as indicated by higher grade point averages, and improved scores of the formative assessments. The regression analysis results showed that AI (personalized learning) and learning analytics (β = .41; p < .001) were among its most important models that were significant in predicting the variable of academic achievement; the analysis of correlation results showed that there was a good positive correlation between the variables of engagement and academic achievement (r = .63, p < .01). The qualitative results corroborated this as students felt there was a significant increase in motivation, personalized feedback, and adaptive learning pathways as being the most beneficial aspects of the AI integration. Issues were identified that could interfere in successful implementation: Data privacy concerns, uneven access to technology and insufficient faculty preparation. According to this study, personalized learning & learning analytics with AI have proved to be significant drivers of student engagement & academic performance, provided that the ethical issues and infrastructure are sorted out and this may influence the way research integrity and rigor will be upheld within Higher Education Institutions. The findings herein expand the conversation taking place around the role of AI in education and highlight the need to embed learning analytics in a learning design practice, to create inclusive and effective learning spaces.