Education on climate change mitigation and artificial intelligence: Critical and empirical analysis

Artificial Intelligence (AI) presents a range of new threats and opportunities across numerous critical public sectors, including education. While the aspiration to develop a genuinely human-centric AI may appear idealistic, increasing resources are being devoted to this goal. Another globally significant issue is climate change mitigation, yet most university curricula inadequately address – or entirely omit – this topic. This article critically explores the risks and potentials of AI in education on climate change mitigation, focusing on the extent to which AI tools in this domain are cross-disciplinary, accessible and impactful. To evaluate AI’s role in education, learning, and teaching methods (AIED), the author employed two primary methodological approaches, placing the focus of scientific interest on the concepts related to interactions between AIED systems and human decision-makers, as well as on their applicability in education on climate change mitigation. First, a critical review of the existing scientific literature on AIED was conducted, focusing on its relationship with educational programs and methods dedicated to climate change mitigation; when it comes to the place of AI in educational ecosystem, the recent studies have expressed the most substantial criticism regarding the issues of major economic and, to a lesser extent, political reasons that have led to its introduction. Second, the evaluation of the impact of AI in education on climate change mitigation was based on two adapted questionnaires, while descriptive and inferential statistical methods were used to process the gathered data. The findings indicate that AI-assisted educational tools specifically tailored to climate change mitigation are significantly more effective in achieving pedagogical objectives than general AI-based tools, allowing their wider use in educational practice. However, challenges remain regarding risks and accessibility, limiting their broader adoption and impact.