Exploring the role of a chatbot in supporting self-regulated learning (SRL) among students with low and high SRL skills

Abstract

With the rise of artificial intelligence (AI) technology, the use of chatbots for providing personalized adaptive learning is increasing. Self-regulated learning (SRL) is a key component of such tools. This study analyses a chatbot called PM Tutor to understand how the differences in the SRL skills amongst the students unfold during their interactions with the chatbot and how these students perceive these interactions. This study employed a combination of data collection methods, including the Metacognitive Awareness Inventory (MAI) questionnaire, log data of the learners' interactions with the chatbot, and retrospective interviews, to understand the nature of these interactions from SRL perspective. The findings revealed that the low SRL group showed higher engagement levels than the high SRL group. However, there was no significant association between high/low SRL and the frequency or sequence of actions when students interacted with the chatbot. During the chatbot interactions, self-assessment and feedback primarily facilitated SRL processes for both groups. The difference in engagement with the chatbot between the two groups could be attributed to the contextual factors as evidenced by students’ perceptions of the chatbot interactions: for example, limited AI capabilities, limited content formats, and usage of the chatbot primarily for revision. Given the context-sensitive nature of SRL, there is potential to enhance the chatbot prototype by addressing these contextual factors. Future research should also seek to evaluate the chatbot’s impact on SRL by collecting data on strategic interactions. Although the chatbot facilitated some SRL processes during revision, further development could better support SRL by establishing clear links between SRL processes and the chatbot’s affordances. This study informs the future design of AI-based educational tools for SRL.