Sustainable AI Regulation

Abstract

Current proposals for AI regulation, in the EU and beyond, aim to spur AI that is trustworthy (e.g., AI Act) and accountable (e.g., AI Liability) What is missing, however, is a robust regulatory discourse and roadmap to make AI, and technology more broadly, environmentally sustainable. This paper aims to take first steps to fill this gap.

In computer science, AI and technology more generally are increasingly recognised as important contributors to climate change–and with good reason: Current estimates show that information and communication technology (ICT) contributes up to 3.9% of global greenhouse gas (GHG) emissions compared to roughly 2.5% for global air travel. The carbon footprint of machine learning more specifically has skyrocketed over the last years. Water consumption is another crucial factor. Regarding both energy and water, AI training is particularly resource intensive, and even more so with large generative AI models such as ChatGPT or GPT-4.

However, questions of climate change and sustainability still occupy a significant blind spot in AI regulation. This article is the first to map out the impact of sustainability considerations on existing and planned technology regulation, such as environmental law, the GDPR and the AI Act. EU environmental law does not, currently, address GHG emissions or water consumption of AI and wider IT infrastructure in a direct and adequate way. However, I posit that, while the GDPR contains references to collective interests and third parties that could be harnessed for an interpretation in line with sustainability goals, existing law could go even further by limiting some of the individual rights we have come to take for granted. For example, individual right to erasure should be balanced against the collective interest in mitigating climate change if its exercise entails significant environmental costs (e.g., retraining a large AI model).

In a second step, the Article suggests concrete policy measures to further align AI and technology regulation with environmental sustainability. In this vein, transparency mechanisms, such as the disclosure of the GHG footprint contemplated under the EU AI Act, could be helpful. Given the well-known limitations of disclosure, however, effective regulation needs to go beyond transparency. Hence, in this paper, I suggest a mix of strategies to work toward sustainable AI regulation: co-regulation; sustainability by design; restrictions on training data; and consumption caps, including the integration into the EU Emissions Trading Regime.

This regulatory toolkit may then, in a third step, serve as a blueprint for other information technologies and infrastructures facing significant sustainability challenges due to their high GHG emissions, for example: blockchain (e.g., bitcoin) Metaverse applications and data centres. The regulatory toolbox described above, from transparency to sustainability assessments and hard consumption caps, can and must be flexibly adapted to these other areas of technology law. Such a comprehensive framework is needed to simultaneously tackle and coordinate the crucial twin transformations of our time: digitisation and climate change mitigation.

Keywords: sustainability, artificial intelligence, AI, ml, machine learning, sustainable AI, climate change, GHG emissions, blockchain

JEL Classification: K00, K32