The Problem of Sustainable AI

A Critical Assessment of an Emerging Phenomenon

Abstract

Recently, the notion of “sustainable Artificial Intelligence (AI)” has gained traction. The contention is that AI technologies hold promise for addressing climate challenges by providing sustainable solutions. In that way, sustainable AI is supposed to harness AI’s capabilities while upholding ethical standards and minimizing its resources use, such as its carbon footprint. In answer to this recent trend, this paper critically questions the very conception of sustainable AI. Drawing on philosophy of technology and critical materialist thinking, it aims to uncover the dominant interests and hegemonic narratives driving sustainable AI developments. The paper begins by outlining the concept of sustainable AI. It then explores the hegemonic power structures and socio-economic dynamics behind AI technologies. Concretely, I show how the promises of sustainable AI largely rely on narratives of efficiency and progress, and work by invoking myths and images of a super-intelligence saving humanity. Following this, I highlight that sustainable AI is the technical solution to the climate crisis from a techno-solutionist vantage point simply reproducing the status quo. The enthusiasm for sustainable AI primarily serves hegemonic interests, rather than genuinely aiming for resource-friendly and ethical solutions. The paper concludes with the observation that if we want true climate action, sustainable AI is not the way to go.