Pendidikan editorial
Open AccessOA2024
Empowering student self-regulated learning and science education through ChatGPT: A pioneering pilot study
A 3-week pilot study with 74 Secondary 4 students in Hong Kong found that a ChatGPT-enhanced chatbot (SRLbot) improved science knowledge, behavioural engagement, and motivation compared to a rule-based chatbot (Nemobot). The number of interactions significantly predicted variations in self-regulated learning (SRL).
D. Ng; C. Tan; J. Leungยท British Journal of Educational Technologyยท 2024ยท DOI 10.1111/bjet.13454
The core problem
Self-regulated learning (SRL) is a critical skill for students to proactively manage their learning in digital environments. Recent advances in artificial intelligence (AI) have opened new possibilities for supporting SRL through personalized guidance and feedback. Generative AI technologies, such as ChatGPT, have shown potential to enhance learning by adapting to individual needs. This study addresses the gap in empirical evidence on the effectiveness of generative AI in promoting SRL among secondary students, particularly in science education. The research compares a generative AI-based chatbot (SRLbot) with a rule-based chatbot (Nemobot) in a 3-week science learning experience involving 74 Secondary 4 students in Hong Kong. The aim is to examine how ChatGPT can empower students' SRL, science knowledge, engagement, and motivation.
Innovation
The study employed a quasi-experimental design with 74 Secondary 4 students from Hong Kong. Participants were divided into an experimental group (using SRLbot) and a control group (using Nemobot). SRLbot is a generative AI-based chatbot powered by ChatGPT, designed to maintain regular study habits and facilitate SRL through personalized recommendations and flexible responses. Nemobot is a rule-based AI chatbot providing predetermined responses. Over three weeks, students engaged in science learning activities with their respective chatbots. Data were collected on science knowledge, behavioural engagement, motivation, and SRL. Quantile regression analysis was used to examine the relationship between the number of interactions and SRL variations. Student feedback on learning challenges, psychological support, and self-regulation behaviours was also gathered to gain insights into their experiences.
Introduction
Self-regulated learning (SRL) is a critical skill for students to proactively manage their learning in digital environments. Recent advances in artificial intelligence (AI) have opened new possibilities for supporting SRL through personalized guidance and feedback. Generative AI technologies, such as ChatGPT, have shown potential to enhance learning by adapting to individual needs. This study addresses the gap in empirical evidence on the effectiveness of generative AI in promoting SRL among secondary students, particularly in science education. The research compares a generative AI-based chatbot (SRLbot) with a rule-based chatbot (Nemobot) in a 3-week science learning experience involving 74 Secondary 4 students in Hong Kong. The aim is to examine how ChatGPT can empower students' SRL, science knowledge, engagement, and motivation.
Why it matters
The findings provide valuable evidence for AI developers and educators to consider generative AI settings and chatbot design to facilitate greater success in online science learning. The adaptability and personalized approach of SRLbot were key factors in its effectiveness, as it could respond to individual students' learning and SRL scenarios in real-time. This contrasts with rule-based chatbots, which lack such flexibility. The study suggests that educators should consider the potential of ChatGPT and other generative AI technologies to support student learning and SRL. However, it also highlights the need for awareness of AI limitations and the importance of equipping teachers and students with AI competencies to use these tools appropriately. The implications extend to policy, emphasizing the integration of AI in education to foster SRL and sustained learning habits. Future research should explore long-term effects and broader applications across subjects and student populations.
Who should read this
Pembaca Research Digest dan praktisi rumpun ini.
Opening member contentโฆ