FROM LECTURER TO CHATBOT: EVALUATING A HYBRID TEACHING MODEL

Abstract

Emergence of generative artificial intelligence (GenAI) is revolutionizing teaching and evaluation in higher education. Early adopters have already demonstrated how large language model (LLM) conversational agents can serve as on-demand tutors, yet empirical evidence regarding their effectiveness in facilitating conceptual learning in non-computational domains is scant. Based on constructivist learning theory, this paper presents a two-year quasi-experimental design carried out at the joint undergraduate business program of SKOLKOVO and MIPT (Moscow). The class of 2023 (n = 42) engaged in a traditional lecture-based Immersion in Human Behavioral Biology. By contrast, students in the 2024 group (N = 42) joined problem-centered sessions facilitated by a human teacher and GPT-4-based chatbot co-facilitator (“Robert”). The post-course FG transcripts were incorporated into and contrasted with expert corpora. Students from the AI-assisted group significantly decreased the semantic distance to disciplinary texts by 60 per cent, suggesting better acquisition of domain-specific language in the absence of a subject-matter lecturer. Satisfaction was generally higher for the 2024 class than the 2023 group. However, qualitative feedback revealed new concerns, such as anxiety regarding the loss of epistemic authority, heightened cognitive load and ambiguity in navigating AI-generated content. Despite deeper conceptual engagement, students reported challenges in interpreting chatbot responses and a need for clearer guidance. These findings suggest that an AI-mediated constructivist model can enhance learning outcomes but must be carefully designed to address motivational, cognitive, and epistemic challenges introduced by the integration of GenAI into the class.

Ähnliche Arbeiten