AI-Augmented Blended Learning and Student Learning Outcomes: Evidence from Higher Education 2023-2025
Abstract
Since 2023, the fast integration of AI-based tools in higher education has transformed blended learning approaches; however, there is little evidence regarding their effects on learning outcomes. The current research aims to investigate the effects of AI-based blended courses on academic performance, engagement, and critical thinking skills of 4,218 students from 62 universities in North America, Europe, and Asia from 2023 to 2025. Through a quasi-experimental design, the effects of these AI-based courses will be compared to those of blended courses without AI assistance, which were taken by 3,904 students. The tools used include feedback generators, AI-based tutors, and content generators. Learning outcomes were assessed through standard tests, class grades, and validated scales for cognitive engagement and self-regulated learning. According to results, AI-empowered classes enhanced average learning outcomes by 12.4%, mostly in low-performers' outcomes, reducing the performance gap by 18%. AI-generated feedback reduced the time required to grade assignments by 37% and increased the frequency of feedback to daily basis from weekly. However, the impact was conditionally related to training of instructors on AI pedagogy as only those classes in which faculty members had at least 15 hours of training on AI pedagogy demonstrated statistically significant effect. Qualitative interview data showed worries about over-reliance on AI and academic honesty issues, but these problems were minimized when there were clear guidelines for use of AI. The results of the study also show that AI tools are more beneficial for formative assessments and personalized practice than for collaborative activities in which profound peer interactions are required. The findings of the study demonstrate that if used properly, AI tools can improve blended learning outcomes with appropriate faculty development.