AI Tutoring Outperforms Active Learning
This week I would like to share a preprint article showcasing how GenAI can support active learning. The article entitled, “AI Tutoring Outperforms Active Learning” by Kestin et al. (2025) examined learning outcomes for students in a large physics course in the fall or 2023 who worked with a custom-designed AI chatbot. When compared with an “active learning” classroom in which students learn in groups, the AI-supported version proved to be more effective. The Harvard authors found that students learned more than twice as much in less time when using an AI tutor, compared with the active learning class. Students also reported that they felt more engaged and motivated. The AI tutor was also able to support natural differences we often find in our classes between students who have a strong background in the content vs. those who lack sufficient background and thus struggle to keep up.
By integrating the best current GenAI technology with research-based pedagogy, the authors believe they have created a learning tool that facilitates active learning, manages cognitive load, scaffolds content, provides timely-targeted feedback, and allows for self pacing. They believe that by attending to foundational learning theory (information processing, metacognition, self-efficacy, self-regulated learning) and research-based teaching methods (sequencing, pacing, etc.) they have provided students with tools that support meaningful learning.
They summarize their findings with the following, which align with decades of research on teaching and learning:
“Our results show that, with today’s GenAI, pedagogical best practices must be explicitly and carefully built into each application. Instructors should avoid using GenAI in situations where students are likely to use it as a crutch to circumvent critical thinking. We advise against the notion that GenAI, solely due to its efficacy in enhancing teaching and learning, should entirely supplant traditional instructional methods. Our demonstration illustrates how GenAI can bolster learning beyond the confines of the classroom. We advocate harnessing this capability to enable instructors to use in-class sessions for activities that foster advanced cognitive skills such as content synthesis.”
References
Kestin, G., Miller, K., & Klales, A. (2025). AI tutoring outperforms active learning, PREPRINT (Version 1) available at Research Square [https://doi.org/10.21203/rs.3.rs-4243877/v1]
Commentaires