As part of the "Digital Economy and the Environment" lecture series, Professor Costa Samaras of Carnegie Mellon University will speak in person at the Center for Industrial Ecology. He will address how the use of Artificial Intelligence enables both opportunities and risks for the energy system and for broader climate mitigation and climate resilience efforts. The rapid evolution of AI capabilities and the corresponding growth in electricity demand, requires robust public policy to ensure electricity for communities is clean, affordable, reliable, and equitable. In addition, a systems understanding of how AI and emerging technologies affect broader energy use and greenhouse gas emissions across society is essential for decision-makers to be able to protect the public, advance innovation, maximize the benefits, and minimize the risks. Using empirically collected power data from H100 DGX server nodes during AI model training, this presentation first outlines a statistical model that predicts AI energy use in data centers. Next the presentation outlines policy options for reinvestment in the power sector – a Grid New Deal – to ensure AI energy use and electrification is good for communities. Finally, the presentation outlines the pathways for AI energy use and climate impacts beyond data centers, and discusses policy pathways that enable climate safe futures under deep uncertainty.
The talk is free and open to all. Please join us in Burke Auditorium in Kroon Hall or virtually Zoom.
