NIST eyes closing data gaps for measuring and limiting AI energy impacts

The absence of data is hindering efforts to both measure and limit the energy and environmental impacts of artificial intelligence, according to researchers on a panel discussion that concluded a National Institute of Standards and Technology symposium on AI innovation.

“But one of the challenges we're facing with AI right now is that there's no agreed upon standard” on how and what to measure when studying environmental impacts, said Arti Garg, chief strategist at Hewlett Packard Enterprise, during the Sept....