ADAPTIVE AI CONVERSATIONAL AGENTS FOR UNIVERSITY-LEVEL LANGUAGE LEARNING: ENHANCING CRITICAL THINKING, INTERCULTURAL COMPETENCE, AND ACADEMIC WRITING SKILLS

Abstract

This study investigates the impact of Adaptive AI Conversational Agents (AACAs) on the development of critical thinking, intercultural competence, and academic writing among university-level language learners, with learner motivation examined as a mediator and cognitive load as a moderator. Guided by Sociocultural Theory, Self-Determination Theory, and Cognitive Load Theory, a quasi-experimental design was employed over a 10-week period with 118 participants divided into experimental and control groups. The experimental group engaged in scaffolded, AI-mediated dialogues and writing tasks tailored to individual performance levels, while the control group received traditional instruction. Quantitative data from standardized critical thinking and intercultural competence assessments, as well as analytic writing rubrics, were analyzed using ANCOVA, structural equation modeling, and moderation analysis. Results indicated significant improvements in all three skill domains for the experimental group (p < .001), with motivation fully mediating writing gains and partially mediating critical thinking gains. Cognitive load significantly moderated the intervention’s effectiveness, with optimal outcomes at low to moderate levels of extraneous load. Qualitative analyses of learner journals and focus groups supported these findings, highlighting perceived benefits in scaffolded challenge, cultural authenticity, and writing confidence. The results suggest that AACAs can serve as both cognitive and motivational catalysts when carefully designed to balance challenge and cognitive manageability. These findings contribute to emerging scholarship on AI-assisted higher education and offer practical guidance for integrating AACAs into sustainable, high-impact language learning environments.