Exploring AI-enhanced collaborative learning : socio-emotional and cognitive perspectives of pre-service teachers

Abstract

This thesis explores how Artificial Intelligence (AI), specifically a Metacognitive AI (MAI) agent, influences collaborative learning experiences among pre-service teachers. The study aims to understand how interactions with AI shape cognitive processes, emotional engagement, and overall participation during group tasks. Existing video data from three groups of Finnish pre-service teachers, each interacting with MAI at varying proactivity levels, was re-examined. The video recordings were carefully coded for different types of participation and interactions (cognitive, socio-emotional, and mixed), and interactions with the AI were examined for recurring themes.

Findings indicate high overall participation (89%), though significant differences emerged among groups. Groups 1 and 3 maintained very high levels of collective engagement throughout (98%), while Group 2 exhibits more fluctuation. Interaction types differed between groups, with some focusing more on cognitive tasks and others engaging predominantly with socio-emotional aspects. Seven key themes emerge from human-AI interaction, portraying AI as both a supportive scaffold and a reflective tool for group collaboration.

The study concludes that AI-enhanced collaboration has substantial potential for teacher education by supporting engagement and shared regulation. However, successful integration of AI into collaborative environments depends significantly on its alignment with existing group dynamics and habits. Future AI tools should therefore balance cognitive support with attention to emotional needs and equal participation. However, this research reinforces established theories of learning (SRL, SSRL, HASRL), emphasizing the continued importance of human interaction and dynamics, even as technology advances.