… and frequently has deleterious consequences for our society and the environment. The sustainable and sensible use of AI is everyone’s responsibility, and we should decide together where to deploy AI – and where not.

Whether AI systems that serve us all will be developed in the future is also dependent on an advantageous policy framework. The first Europe-wide obligation to make AI products more sustainable may soon be introduced in the form of the European Union’s AI Act. EU member state governments, however, have not shown much interest in including a requirement that the environmental impacts of AI technologies be measured. Omitting that obligation would be negligent and a failure on the part of European leaders to live up their political responsibilities. It is only possible to know how environmentally damaging these technologies are if comprehensive measurements are undertaken. The problem will not simply go away if we ignore it. Meaningful data would help us better understand the problem and enable us to put more pressure on policymakers. That’s why we need more of it. And that’s why we need to start measuring.

This magazine is an invitation to think more about what concrete steps could be taken to better regulate the development and implementation of AI technologies – through greater transparency regarding their consumption of energy and water, through prohibitions of harmful uses of AI and stronger incentives to improve efficiency.

Dr. Anne Mollen
SustAIn project leader