Between major technological transformation, hype and hypothetical promise, Artificial Intelligence is on everyone's lips. However, its environmental impacts are generally not very visible and often underestimated.
Starting with a historical and technological perspective, this dedicated note explores the direct and indirect effects of generative AI on energy consumption, water, rare metals and equipment obsolescence.
This note also addresses the blind spots: between promises of progress and unsustainable uses, what are the real levers for more sober AI? And above all, can we still steer its development towards models that are compatible with planetary boundaries?
A joint publication with our partners Easyvirt, EVEA, Greenspector, and Toovalu.
