Large language model (LLM)-based artificial intelligence tools, such as ChatGPT, Claude, and Microsoft Co-Pilot, have taken the world by storm, leaving universities playing catch-up to regulate and guide students in the use of these algorithmic aids. Universities recognise the opportunities these tools offer for teaching and learning, as well as the importance of mastering them in future job markets (Digital Education Council, 2025). Yet, careless adoption of these tools carries significant cost and risk at several fronts: first, to civic institutions (Hartzog & Silbey, 2025); second, to natural resources (de Vries-Gao, A. 2025; International Energy Agency, 2025; Jegham et al., 2025); third, to epistemic resources (Dammu et al., 2024; Sommerer, 2024), and fourth, to students’ cognitive (Gerlich, 2025; Lee et al., 2025) and behavioural development (Zhou & Zhang, 2024; Kooli et al., 2025). As such, choosing appropriate use cases for LLM-based AI tools (when, where and how) is an everyday challenge for university teachers and students alike. In the rest of the text, I refer to these tools with the term ‘AI tools’ in the interest of brevity and in concordance with the common practice, despite large language models representing only a small, even if topical, aspect of the vast field of artificial intelligence. 

Building on the concept of digital competence (Ilomäki et al., 2016; Kasperski et al., 2022; Spante et al., 2018; Zhao et al., 2021), my research explores the foundations of responsible large language model use. Specifically, I develop a conceptual framework that foregrounds two key aspects of responsible use. First, the framework foregrounds the development of students’ ability to use AI in a meaningful way, encouraging consideration of the purpose, process and audience of a task alongside its output. Second, the framework supports the development of students’ ability to critically evaluate AI tools and the cost of their use, increasing the awareness of the manifold implications for natural and epistemic sustainability. 

My paper explores a conceptual framework of responsible AI use. On this conceptual analysis, I draw on pedagogical literature on digital and AI competence (e.g., Ilomäki et al., 2016; Kasperski et al., 2022; Spante et al., 2018; Zhao et al., 2021) as well as on a broad range of interdisciplinary research on large language model development, use, and impact, including computer science, environmental studies, and psychology. As a part of this work, I am preparing an action research project as the next empirical research step (Niemi, 2011; Norton, 2009). I aim to have the initial results of this phase of the study by August 2026 and the Emerging Research Conference.

The ubiquity of AI tools is risking the sustainability transition in society (Bush et al., 2025). My responsible AI use framework aims to contribute to a sustainable AI practice in the short term by providing educators with a tool for teaching interventions and, in the long term, with rigorous empirical research that informs education and has the potential to influence policy and educational practice.

Bush, A., Aksoy, M., Pauly, M., Ontrup, G. (2025) Choosing a model, shaping a future: Comparing LLM perspectives on sustainability and its relationship with AI. arXiv:2505.1443. Dammu, P.P.S., Jung, H., Singh, A., Choudhury, M., Mitra, T. (2024) “They are uncultured”: Unveiling covert harms and social threats in LLM generated conversations. arXiv:2405.05378v1. de Vries-Gao, A. (2025) Artificial intelligence: Supply chain constraints and energy implications, Joule. Digital Education Council. (2025). AI in the Workplace 2025. Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6 Hartzog, W., & Silbey, J. M. (2025). How AI Destroys Institutions. Boston Univ. School of Law Research Paper No. 5870623. Ilomäki, L., Paavola, S., Lakkala, M., & Kantosalo, A. (2016). Digital competence–an emergent boundary concept for policy and educational research. Education and information technologies, 21(3), 655-679. International Energy Agency. 2025. Energy and AI. IEA Publications. Jegham, N., Elmoubarki, L., Abdelatti, M., Koh, C. Y., Hendawi, A. (2025). How hungry is AI? Benchmarking energy, water, and carbon footprint of LLM inference. arXiv:2505.09598v6. Kasperski, R., Blau, I., & Ben-Yehudah, G. (2022). Teaching digital literacy: Are teachers’ perspectives consistent with actual pedagogy? Technology, Pedagogy and Education, 31(5), 615-635. Kooli, C., Kooli, Y., & Kooli, E. (2025). Generative artificial intelligence addiction syndrome: A new behavioral disorder? Asian Journal of Psychiatry, 107, 104476. Lee, H. P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025, April). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. In Proceedings of the 2025 CHI conference on human factors in computing systems (pp. 1-22). Niemi, R. (2019). Five approaches to pedagogical action research. Educational Action Research, 27(5), 651-666. Norton, L. (2009). Action research in teaching and learning: A practical guide to conducting pedagogical research in universities. Routledge. Sommerer, T. (2025). Baudrillard and the Dead Internet Theory. Revisiting Baudrillard’s (dis) trust in Artificial Intelligence. Philosophy & Technology, 38(2), 54. Spante, M., Hashemi, S. S., Lundin, M., & Algers, A. (2018). Digital competence and digital literacy in higher education research: Systematic review of concept use. Cogent Education, 5(1), 1519143. Zhao, Y., Llorente, A. M. P., & Gómez, M. C. S. (2021). Digital competence in higher education research: A systematic literature review. Computers & education, 168, 104212. Zhou, T., & Zhang, C. (2024). Examining generative AI user addiction from a CAC perspective. Technology in Society, 78, 102653.