New DOE Workplace Coordinating AI
Two visitor audio system from DOE’s new Workplace of Crucial and Rising Applied sciences (CET) joined the webinar to speak about their workplace’s function working with DOE and its nationwide laboratories, which entails figuring out potential makes use of for AI within the vitality sector and guaranteeing these makes use of are developed safely. Michael Mazur, a CET coverage advisor, mentioned the workplace was established to assist implement the 2023 Government Order on the Protected, Safe, and Reliable Improvement and Use of Synthetic Intelligence.
“In parallel with selling the advantages of AI, we’re actually targeted on managing the dangers,” mentioned Mazur. “We’re deploying AI testbeds, or locations the place we will strive it out—whether or not it’s new {hardware}, new software program, or different new ideas. And we’re asking ourselves the exhausting questions: When an AI mannequin is on the market, what’s it able to doing? So, we’re doing all these issues in protecting environments so we will quantify these dangers, after which use that to tell the long run broader regulatory area.”
Charles Yang, one other CET coverage advisor, offered extra specifics on how the workplace is assessing ways in which AI can help a contemporary energy grid—from grid planning and operations to reliability and resilience issues. He additionally shared a latest DOE report that discusses how AI will help guarantee clear vitality can meet the problem of powering AI itself, a rising concern within the sector.
Yang concluded by speaking about DOE’s latest announcement of Frontiers in AI for Science, Safety, and Know-how, or FASST. “The objective of FASST is to sort out a whole lot of the problems we see on this area,” mentioned Yang. “So, that is each advancing scientific discovery and nationwide safety, but in addition addressing the vitality problem by way of continued improvements and vitality effectivity, constructing out that AI workforce, and eventually, creating the technical experience that we want for AI governance.”
