Comparison of Psychological Safety in Learning Environments between AI-Based Classes and Conventional Classes

Abstract

This study compares psychological safety between Artificial-Intelligence-based learning environments and conventional classrooms by examining both social and cognitive dimensions that influence learners’ comfort. The study was conducted because although both AI-based and face-to-face instruction have been proven to enhance learning outcomes, they have not been compared specifically in relation to psychological safety. A comparative qualitative method was used, involving in-depth interviews and document analysis, followed by thematic analysis to capture recurring experience patterns. Findings reveal that AI-based classrooms enhance psychological safety through evaluative privacy, personalized learning, reduced academic anxiety, and lower cognitive load, whereas conventional classrooms strengthen psychological safety through social support, emotional closeness, and interpersonal engagement. However, both have limitations: AI-based learning may reduce social connectedness, while conventional learning is vulnerable to performance anxiety. The study concludes that optimal psychological safety is achieved through an integrated hybrid learning model that combines the strengths of AI and human interaction.