Abstract
In recent years, the notion of 'sustainable AI' has emerged as a new topic within the wider debate on artificial intelligence (AI). Although sustainability is usually understood as having three dimensions - the environment, society, and the economy - the debate on sustainable AI, so far, is characterized by a narrow focus on the environmental sustainability of AI. In this article, it is argued that the debate on sustainable AI should not only be about AI's environmental costs, but also incorporate social and economic concerns. More precisely, the article shows that AI's environmental impact is often connected to important social and economic issues. Through an analysis of existing literature on the sustainability of AI, it is demonstrated that ethical and philosophical arguments about the importance of considering the environmental costs of AI apply just as well to the social and economic dimensions of AI development. The aim of this article is thus to present a broader understanding of sustainable AI; one that includes all three pillars of sustainability and acknowledges the interdependence between AI's environmental, social and economic costs.