Evidence really mounting on GenAI tutors performing as well as or better than other teaching methods. In this Harvard study AI Tutoring outperformed even active learning (Kestin et al. 2024).
Recent breakthroughs in generative AI (GAI) are set to transform education as we know it. But we're still figuring out the best ways to use this new technology and how effective it really is. This study is particularly interesting, as it compares University students learning with an AI-powered tutor to those in a traditional active learning class. They made sure the AI tutor used the same teaching methods as the regular lectures. And guess what?
Results
Compared to those in the regular class, students using the AI tutor learned:
- more than twice as much
- in less time
If this continues, these findings show that AI tutoring could massively boost learning, making a strong case for adopting it widely in education.
The authors note that passive lectures are the least effective, being passive, going at the wrong pace for many learners, with no feedback and little engagement. Even active learning is never one-on-one and therefore suffers from the same problems as passive lectures.
Motivation
Improved levels were found on other qualitative measures:
- Engagement
- Motivation
- Enjoyment
- Growth mindset
This matters, especially ‘motivation’, where I have seen significant impact in many contexts. This, I some ways is more important the younger the learner.
Conclusion
Little, other than the flawed Bloom 2 Sigma paper, is being seen in presentations on efficacy. I have been presenting this and other papers, to show that we are quickly gathering evidence of GenAIs effectiveness in learning. The hypothesis, that GenAI is a powerful learning tool seems obvious when one looks at the astonishing use by so many in so many contexts, school, University and workplace. Academic publishing lags long behind reality. This is a welcome addition to the evidence.
Paper
https://www.researchsquare.com/article/rs-4243877/v1
No comments:
Post a Comment