Alexandra Luccioni, a collaborator on that study, says that keeping an eye on the emissions level of AI is crucial for the near future. They developed an emissions calculator to estimate the energy use, and the concomitant environmental impact, of training ML models. The emissions incurred in the training of a neural network model, they found, are related to the location of the training server and the energy grid it uses, the length of the training procedure, and the hardware on which the training takes place. Alexandre Lacoste and colleagues worked on a study to make quantifying the carbon cost of ML easier for researchers.
