Teacher’s mentor role in primary students’ CHATGPT-integrated blended learning

Abstract

Artificial intelligence is gradually permeating myriad facets of human existence, leaving no domain untouched, including the educational system. Educators approach such integration with circumspection, primarily apprehensive that the nascent digital generation of Centennials may forfeit critical thinking abilities and autonomous information analysis skills, while concurrently becoming excessively reliant on technology, thereby forfeiting the capacity to engage in traditional learning methodologies. This study elucidates the role of the instructor as a mentor within blended learning paradigms to cultivate logical capacity in elementary school pupils, fostering a propensity for self-directed learning coupled with a critical assimilation of information furnished by a chatbot underpinned by an expansive language model.

This experimentation seeks to investigate the impact of third-grade students utilizing the ChatGPT bot as a virtual guide within a blended learning framework on their academic performance and attitudes toward artificial intelligence in education. The methodology encapsulated 43 primary school students belonging to a reference group and a treatment group. The latter deployed ChatGPT to check and solve assigned logic riddles over a fortnight. Statistical strategy incorporated the D’Agostino-Pearson test, paired t-test, and intergroup t-test. The outcomes divulged that the treated cohort exhibited a significant augmentation in learning attainment, with their post-experimental points surpassing controls by 10.6%. Subjects who exploited the chatbot agent as an ancillary scholarly means for autonomous learning appraised its capacity to dope out logical conundrums favorably. However, their perception of the language model fluctuated contingent upon its power to yield accurate responses. Specifically, the revelation of errors by the students diminished their interest in ChatGPT. Pedagogically, the model’s inaccuracies fostered the gain of critical thinking among the partakers and heightened their engagement in the learning routine.

In sum, the study unveils the potential of leveraging ChatGPT as a virtual mentor to enhance the efficacy of self-directed learning among elementary school students. When undertaking homework assignments, children can solicit the chatbot to review their work and harness the language model’s capabilities as an educational tool capable of rectifying incorrect answers, resolving uncertainties, and guiding learners toward a more profound comprehension of academic subjects, thereby facilitating more robust and enduring knowledge retention.

Key words: blended learning, students, mentor, artificial intelligence, chatbot, ChatGPT, digital education.