Artificial Intelligence-Based Blended Learning: A Mixed-Methods Study

This study aims to investigate the effects of using artificial intelligence (AI) tools in educational environments. A mixed method was used to investigate the effects of an AI application (ChatGPT) used in a blended environment on achievement scores and students’ learning experiences. The experimental group was given artificial intelligence-based blended learning (AIBL), one control group was given face-to-face learning (F2F), and the other control group was given online learning (OL). The quantitative findings indicated a statistically significant difference among the three instructional conditions in post-test achievement scores, F (2,80) = 7.33, p = .001, η ² = .16. Post hoc comparisons indicated that the AIBL ( M = 37.58, SD = 10.40) scored significantly higher than the OL ( M = 28.02, SD = 5.90), whereas no statistically significant difference was found between the AIBL and the F2F ( M = 35.40, SD = 10.92; p = .667). These findings suggest that AIBL may offer an advantage over fully OL, while producing achievement outcomes comparable to F2F instruction. Qualitative findings were analysed in the themes of experience with ChatGPT in AIBL, perceived benefits, perceived limitations, and learning support. Preservice teachers believe that the use of AI applications in the educational environment has both positive and negative effects on their learning. ChatGPT’s ability to facilitate students’ access to summarised and focused information was rated positively. Preservice teachers have reported that they sometimes perceived ChatGPT’s answers and operations as incorrect and that this is the main adverse effect of AI. This study provides important insights that will guide the effective use of AI applications in educational environments.