The Environmental Cost of Intelligence: Corporate Obligations and a Governance Framework for Sustainable AI
DOI:
https://doi.org/10.51325/6b52xt27Keywords:
Green AI, Sustainable AI, AI Environmental Footprint, Carbon Emissions, Corporate Environmental Responsibility, AI GovernanceAbstract
The field of artificial intelligence has advanced incredibly fast in recent years with an increasing focus on larger language models, generative technologies, and infrastructure of deep learning. Such development has resulted in a substantial environmental cost that has been relatively ignored – the energy, water, and carbon emissions related to training, fine-tuning, and industrial-scale operation of AI. Training one single large language model is estimated to generate as many carbon dioxide emissions as several cars throughout their entire lifetime. Data centers used for AI operations are predicted to double their energy consumption by 2026, and the amount of water required to cool down AI-related equipment is becoming a limitation in places that are already facing water shortages. Research concerning these environmental impacts of AI has flourished, while corporate disclosure of its environmental footprint is extremely rare and highly inconsistent. In this paper, the Sustainable AI Governance Framework (SAGF) is developed as a conceptual model which defines environmental obligations for corporations during the whole AI lifecycle, from the development and training stage, through its deployment, maintenance, and decommissioning stages, and introduces governance approaches on three different levels – the firm level, industry level, and regulatory level. The SAGF uses principles of corporate environmental responsibility theory (Bansal & Roth, 2000; Schaltegger & Burritt, 2018), polluter pays principle, and recent developments regarding the environmental impact of AI (Patterson et al., 2022; Lottick et al., 2019; Strubell et al., 2019) to prove that the developers and users of AI technology have environmental obligations that are not able to be shifted elsewhere.
References
Bansal, P., & Roth, K. (2000). Why companies go green: A model of ecological responsiveness. Academy of Management Journal, 43(4), 717 to 736. https://doi.org/10.2307/1556363 DOI: https://doi.org/10.2307/1556363
European Parliament. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union.
GHG Protocol. (2015). Corporate accounting and reporting standard (Revised ed.). World Resources Institute & World Business Council for Sustainable Development.
Heede, R. (2019). Carbon majors: Updating activity and emissions data for the world's largest oil and gas companies. Climate Accountability Institute.
IEA. (2024). Electricity 2024: Analysis and forecast to 2026. International Energy Agency.
Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 389 to 399. https://doi.org/10.1038/s42256-019-0088-2 DOI: https://doi.org/10.1038/s42256-019-0088-2
Li, P., Yang, J., Islam, M. A., & Ren, S. (2023). Making AI less 'thirsty': Uncovering and addressing the secret water footprint of AI models. arXiv preprint arXiv:2304.03271.
Lottick, K., Susai, S., Friedler, S. A., & Wilson, J. P. (2019). Energy usage reports: Environmental awareness as part of algorithmic accountability. NeurIPS Workshop on Tackling Climate Change with Machine Learning. arXiv preprint arXiv:1911.08354. https://doi.org/10.48550/arXiv.1911.08354
Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019). Model cards for model reporting. Proceedings of the Conference on Fairness, Accountability, and Transparency, 220 to 229. https://doi.org/10.1145/3287560.3287596 DOI: https://doi.org/10.1145/3287560.3287596
OECD. (2008). The polluter pays principle: Definition, analysis, implementation. OECD Publishing. https://doi.org/10.1787/9789264044845-en
Patterson, D., Gonzalez, J., Hölzle, U., Le, Q., Liang, C., Munguia, L.-M., Rothchild, D., So, D. R., Texier, M., & Dean, J. (2022). The carbon footprint of machine learning training will plateau, then shrink. Computer, 55(7), 18 to 28. https://doi.org/10.1109/MC.2022.3148714 DOI: https://doi.org/10.1109/MC.2022.3148714
Schaltegger, S., & Burritt, R. (2018). Business cases and corporate engagement with sustainability: Differentiating ethical motivations. Journal of Business Ethics, 147(2), 241 to 259. https://doi.org/10.1007/s10551-015-2938-0 DOI: https://doi.org/10.1007/s10551-015-2938-0
Strubell, E., Ganesh, A., & McCallum, A. (2019). Energy and policy considerations for deep learning in NLP. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 3645 to 3650. https://doi.org/10.18653/v1/P19-1355 DOI: https://doi.org/10.18653/v1/P19-1355
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Naglaa Ibrahim Khamis (Author)
This work is licensed under a Creative Commons Attribution 4.0 International License.
