Cities can green AI procurement through environmental criteria

Municipalities increasingly face pressure to adopt artificial intelligence while simultaneously meeting climate and emissions reduction targets. As residents demand climate action and public sector staff worry about how AI infrastructure will affect both environmental goals and IT budgets, there is a clear need for a more environmentally sustainable approach to AI procurement and use.

A new handbook addresses this gap by providing practical guidance on how to incorporate environmental performance criteria into public AI procurement. The guide is grounded in substantial field research: it draws on 10 months of fieldwork conducted with municipal governments, complemented by five months of co-design interviews and focus groups involving city staff and green software practitioners. This collaborative development process ensured that the recommendations are not only scientifically sound but also feasible within the constraints of real government operations.

The handbook's approach focuses on making environmental responsibility actionable at the municipal level. Rather than proposing wholesale changes to procurement systems, it introduces entry-level, low-effort actions that can be integrated into existing procurement processes. These practical steps are designed to help agencies optimize sustainability while preserving the quality of service that residents depend on. The guide is intended for a broad audience—anyone involved in AI adoption in an organization—recognizing that environmental performance criteria are relevant across public sector decision-making, not just at the senior policy level.

Municipal governments face a growing tension as they pursue both digital innovation and climate commitments. Residents and public sector staff increasingly scrutinize the environmental cost of AI infrastructure—not only in terms of direct emissions, but also operational expenses that strain IT budgets. The handbook addresses this pressure by demonstrating that environmental responsibility and procurement quality are not mutually exclusive; instead, it positions sustainability as an integral part of responsible AI governance.

The development process itself reflects the practical orientation of the guide. Rather than relying solely on academic research or vendor input, the authors spent 10 months conducting fieldwork and five months in co-design interviews and focus groups with the very people who implement these decisions—city staff and green software practitioners. This ground-level engagement ensures that the suggested actions are feasible within existing municipal workflows. By framing environmental criteria as low-effort, entry-level additions to standard procurement processes, the handbook lowers the barrier to adoption and signals that greener AI is achievable without wholesale process overhaul.

No comments yet. Be the first to share your thoughts!