Author Information
Miguel A. Varela, Universitat Oberta de Catalunya, SpainAbstract
The increasing integration of artificial intelligence (AI) in education has posed a question for scholars: are learners prepared for fully AI instruction without teacher intervention? There has been some research investigating whether AI-supported agents can act as teachers, and some researchers have even raised the question of whether teachers might still be necessary in the future. In this hypothetical case of fully AI-mediated lessons, are younger learners ready to learn solely through AI technology? This study addresses this concern by examining whether high school students can learn Spanish as a foreign language effectively solely with AI and how this compares with traditional teacher-led education. Based on student agency and self-directed learning theories (Candy, 1991; Knowles, 1975), the study employed a crossover design in an international school. Participants worked on two units (reflexive verbs and past tense) under two conditions: teacher-led instruction and autonomous learning with a custom AI-supported app. Data were collected through pre- and posttests, a confidence survey, and semi-structured student interviews. Quantitative findings revealed significant learning gains in both instructional conditions, but there was not statistically significant superiority of either method. Qualitative findings, however, indicated that students continued to value teachers for clarification, feedback, emotional support, and personalized explanations. The results suggest that while AI technologies can effectively support language learning and engagement, students still preferred the human teacher interaction. This study presents original empirical data on AI integration, learner autonomy, and the future of education. Findings support hybrid methodologies that integrate technological agency with teacher support.
Paper Information
Conference: WCSS2026Stream: Education and Social Welfare
This paper is part of the WCSS2026 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window
Comments
Powered by WP LinkPress