Chapter 09 · Sustainability & AISustainability & AI
Sustainable AI
Definition
An umbrella concept covering both the use of AI to advance sustainability goals ('AI for sustainability') and reducing the environmental and social impact of AI itself ('sustainability of AI').
References
This source provides part of the technical or institutional basis for the definition.
This source supports the explanation of how the term is applied, measured or governed in practice.
Overview
What it means
The dual framing comes from van Wynsberghe's 2021 paper, which defined Sustainable AI as change across the entire AI life cycle toward ecological integrity and social justice. The second branch — AI's own footprint and supply chain — has historically received less attention than the first.
How it is used
The term structures policy programmes, research agendas and corporate strategies: application portfolios (climate modelling, biodiversity monitoring) on one side; footprint management, efficient models and responsible hardware sourcing on the other.
Why it matters
Keeping both branches in one term prevents a common sleight of hand: showcasing AI-for-good applications while externalising the costs of the infrastructure underneath. Sustainable AI insists the ledger be read whole.