Work in Progress

Guan, R., Raković, M., Chen, G., & Gašević, D. (2025). How educational chatbots support self-regulated learning? A systematic review of the literature. Education and Information Technologies, 30, 4493–4518. https://doi.org/10.1007/s10639-024-12881-y

Guan et al. (2025) conducted a systematic literature review to examine how educational chatbots contribute to students’ self-regulated learning and academic achievement. The authors employed Winne and Hadwin’s model of self-regulated learning to analyze the studies. This model has four components: understanding the task at hand; setting goals and developing plans; using learning strategies; and adjusting one’s approach to learning in response to feedback. The authors searched for studies published between 2012 and 2023. Ultimately, 27 studies met the authors’ inclusion criteria. While most chatbots help students find resources, use learning strategies, and monitor their progress, few support goal-setting, planning, reflecting on one’s learning process, or adjusting one’s learning methods based on prior experiences. Although many studies reported positive outcomes, several exhibited either mixed results or failed to report statistically significant findings.
This study was transparent with the review process. The authors followed PRISMA guidelines throughout the review process and used the SPIDER framework as a basis for guiding their search for relevant studies. In addition, the authors described how they selected studies for inclusion/exclusion and demonstrated that reviewer agreement occurred regarding study selection. The organization of the results according to the self-regulated learning model made it clear how the use of chatbots supports various learning phases. However, they had some limitations. The review was based on only 27 studies, and the majority of those were with college students. A lot of variability existed in the types of chatbots used, participant characteristics, research methods employed and how performance was assessed. It is therefore impossible for us to definitely say that chatbots will always result in better learning outcomes. In addition, the review did not provide a meta-analysis to determine an overall measure of effect.

I am interested in AI-assisted learning. Few chatbots provide opportunities for learners to plan their own activities and then reflect upon those activities. My interest is in creating a tool that will enable parents, teachers, and other professionals to better understand children’s needs and decide the next appropriate steps. The tool incorporates AI functions that would help users define the problem, evaluate data, set realistic goals, and reflect on whether the action was effective. This article helps me recognize the potential educational value of AI that in supporting users’ thinking, reflection, and decision-making.