Abstract:

In the context of rapidly evolving discussions on sustainability and artificial intelligence (AI), understanding the synergies and tensions between these critical areas is increasingly vital. The growing effort towards AI for sustainability must be complemented by the sustainability of developing and using AI systems. To master the associated challenges and to be able to leverage its potential, we will need experts who possess specific skills to sufficiently understand, analyze, and ultimately shape complex systems soon. Addressing this need requires interdisciplinary teaching and learning offerings that enable students to work independently as well as collectively, network knowledge, and come to informed decisions and solutions to complex problems. Consequently, a new teaching format has been implemented: sustAInability – advanced topics on sustainability and AI. This initiative, a collaborative effort by Technical University of Munich and Hochschule München University of Applied Sciences, funded by Bavarian Institute for Digital Transformation (bidt) for four years, utilizes challenge-based learning to immerse students in real-world problems sourced from industry, academia, and the public sector. The seminar features four intensive sessions exploring the sustainability of AI and AI for sustainability from economic, ecological, and social perspectives, highlighted through expert examples. Students then engage with real-world challenges presented by partners from research, public administration, industry, and civil society, focusing on sustainability and AI. To guide students through the process, various methods including design thinking are incorporated into the seminar. The pressing question is whether, in such a condensed time frame, students can cultivate and enhance their efficacy in sustainability and AI by addressing real-world-challenges?

This paper examines the effectiveness of this methodology in terms of the students' self-efficacy regarding the combination of sustainability and AI, as well as the learners' willingness, abilities, and behavior to contribute to the Sustainable Development Goals (SDGs). Therefore, we conducted a preliminary evaluation of the second cohort during the winter semester of 2023/24, where 20 out of 32 participants responded after completing the seminar. The responses indicate that, through challenge-based learning in our seminar, students’ literacy and self-efficacy in sustainability and AI have improved, regardless of their background or prior knowledge. We conducted a second survey, which is part of a pre- and post-evaluation, aimed at measuring learners' willingness, abilities, and behavior to contribute to the SDGs. 30 out of 46 participants voluntarily took part in the survey. It shows that the designed questionnaire has good potential to improve educators' understanding of Sustainable AI-related abilities, attitudes, and actions of cohorts of students within higher education.

Keywords:

Challenge-based-learning, sustainability and AI, design thinking.