AI-Mediated Learning in Education for Sustainable Development: Supporting Critical Thinking and Responsibility as Action in Complex Learning Contexts

Abstract

The increasing use of Artificial Intelligence (AI) in education presents both opportunities and challenges for developing key competencies within Education for Sustainable Development (ESD), particularly critical thinking and responsibility as action. In line with the goals of ESD and Sustainable Development Goal 4.7 (SDG 4.7), this study explores how AI-mediated learning supports students’ engagement with complex and uncertain knowledge contexts. Grounded in ESD theory and Freire’s concept of praxis (action–reflection–action), the study adopts a Theory-Based Impact Evaluation (TBIE) framework to examine how learning occurs through the activation of mechanisms such as reflection, evaluation, and decision-making. A qualitative classroom-based design was used by employing the Nexgen Classroom platform to facilitate AI-mediated learning tasks. Data were collected through platform interaction logs and student questionnaires, including both closed and open-ended responses. The findings indicate that AI-mediated learning appears to create opportunities for critical thinking and responsibility by engaging students in interpreting, evaluating, and making decisions about AI-generated feedback. However, variations in student engagement suggest that these processes are not consistently activated across all learners. This suggests that while AI-induced uncertainty can stimulate key learning mechanisms, its effectiveness depends on purposeful pedagogical guidance. This study contributes to the understanding of AI in education by conceptualizing AI not simply as a tool, but as a complex learning context that can support the development of ESD competencies when used thoughtfully. It highlights the importance of designing learning environments that encourage active engagement, critical evaluation, and responsible decision-making, thereby supporting the broader goals of sustainable education. Aim: This study aims to explore how AI-mediated learning supports the development of critical thinking and responsibility as action as key ESD competencies. It focuses on how students engage with AI-generated feedback in complex and uncertain learning situations, in line with the goals of SDG 4.7. Theory: The study is grounded in ESD and Freire’s concept of praxis (action–reflection– action). It uses TBIE to understand how learning occurs through mechanisms such as reflection, evaluation, and decision-making. AI is conceptualized as a complex learning environment that introduces uncertainty and multiple perspectives by supporting deeper cognitive engagement. Method: The study uses a qualitative, classroom-based design supported by descriptive data. Data were collected through 1) AI-mediated tasks using the Nexgen Classroom platform 2) Platform interaction data (student responses and engagement) 3) Student questionnaires (closed and open-ended responses) The analysis follows a TBIE approach, where data are examined to identify learning mechanisms (e.g., reflection, evaluation, decision-making) that explain how learning develops. Results: The findings indicate that AI-mediated learning supports critical thinking by requiring students to interpret, compare, and evaluate AI-generated responses. At the same time, responsibility as action develops as students make decisions about whether to accept, modify, or reject AI feedback. However, the results also suggest variation: while some students engage deeply, others struggle with uncertainty or rely less on critical evaluation. The findings indicate that AI can support ESD competencies, if it is combined with structured guidance and pedagogical support.