Timothée Poisot

Although one of the mandates of ecological research is to advance the cause of sustainability, we need to employ simulation and data analysis tools that can incur a massive carbon footprint. Recent calls to rely increasingly on AI, and in particular generative AI, create the very real risk that we will have a disproportionate environmental impact for reasons that rarely go beyond “it is the hyper thing to do right now”.

In a new paper led by former lab member Kari Norman, we discuss the potential for more conscientious computation, where we encourage the selection of methods of both data analysis and simulation that attempt to have the smallest possible environmental footprint.