This week in EDU, we’re thinking about evidence-based research in AI and education.
What does the latest research reveal about the effectiveness of AI in teaching and learning? As studies emerge, they paint a clearer picture of how AI is reshaping education. Evidence-based research is highlighting both the opportunities AI presents for personalized, scalable teaching and the challenges it poses to traditional learning models.
A recent post from
of challenges the idea that there is “absolutely no evidence” of AI’s instructional benefits. It highlights promising areas — like AI tutoring, writing feedback, and hybrid teaching models — while also addressing concerns. This conclusion stands out:The most compelling evidence emerges from hybrid approaches that thoughtfully combine AI capabilities with human instruction. This synthesis appears to leverage the strengths of both: AI’s ability to provide consistent, personalized feedback at scale, and human teachers’ capacity for emotional connection, complex judgment, and holistic understanding.
Here’s the key research
analyzes in depth:“AI Tutoring Outperforms Active Learning”: Kestin et al.’s research at Harvard reveals that AI tutoring can enhance active learning methods, enabling students to learn more than twice as much in less time (study).
“‘ChatGPT is the companion, not enemies’: EFL learners’ perceptions and experiences in using ChatGPT for feedback in writing”: Teng’s research in Macau helps us understand how ChatGPT can assist in the writing development of students learning English as a Foreign Language, improving students’ writing motivation, self-efficacy, engagement, and collaborative writing tendency (study).
“Exploring the Impact of Artificial Intelligence in Teaching and Learning of Science: A Systematic Review of Empirical Research”: Almasri’s systematic review represents a meta-analysis of AI's impact across science education, including how AI can be used for virtual laboratory simulations, tutoring, automated assessment, data visualization, and interactive learning environments (study).
“Improving Student Learning with Hybrid Human-AI Tutoring: A Three-Study Quasi-Experimental Investigation”: Thomas et. al.’s study and quasi-experiment across three different middle schools showed how combining human and AI tutoring increased student proficiency in middle school math (study).
“Critical thinking in the AI era: An exploration of EFL students’ perceptions, benefits, and limitations”: Another study on the impact of AI on students learning English as a Foreign Language, Darwin et. al.’s qualitative investigation focuses on master’s degree students at Indonesian universities and concludes that AI can improve critical thinking skills, with “caveats that require careful management” (study).
Significant gaps remain, particularly in longitudinal studies and research conducted in diverse educational contexts. As we close out 2024, the continued exploration and collaboration within this community are essential to understanding AI’s role in education.
Have new research or perspectives to share?
We want to hear from you in the comments below.
What we’re reading on Substack this week
analyzes what five recent studies say about AI’s impact in the classroom: reflects on how the skills people need to effectively extend their mind apply to generative AI: invites 10 authors to reflect on the future of AI, including , who discusses “The Near Future of AI in Education,” and , who considers “The Seduction of Outsourced Thinking”:AI and EDU insights this week
Using A.I. when your students are not allowed to use it is hypocritical. If students are putting in work and effort, why should teachers be allowed to be lazy and let A.I. grade for them? It gives a sense that, because I am older and I am your superior, I can use A.I. and you can’t, which is not a good example for students. As a teacher, you set the example.
—Aidan at Fountain Valley High School in the New York Times
What Students Are Saying About Teachers Using A.I. to Grade
The trends suggest that students should be learning a collaborative model of software development where a human and an AI assistant work together to generate code. However, there is a larger issue of whether computer science skills as we define them today are even suitable for the future workforce.
—AI Club’s Nisha Talagala in Forbes
How AI Will (or Should) Change Computer Science Education
Just about everyone working in technology whom I interviewed for this article still thinks you should learn to code. But some see a parallel with long division: It’s good to understand how it works. It’s an arguably necessary exercise for learning more advanced mathematics. But on its own, it gets you only so far.
—Reporter Sarah Kessler in the New York Times
Should You Still Learn to Code in an A.I. World?