Who Pays for AI's Power? Carbon Costs and Community Impact

AI's rapid expansion drives large-scale energy use, greenhouse gas emissions, and resource extraction. Training and serving large models consume significant electricity; data center cooling and power provisioning increase local water and land use; hardware production and disposal create upstream mining impacts and e-waste. These environmental effects contribute to climate warming, air and water stress, and broader ecological harm and they are often concentrated near low income communities and communities of color, amplifying local pollution and health risks. This talk will spotlight the scale and geography of AI's environmental footprint: energy and emissions from training and inference, water and land demands for cooling and power, supply chain impacts from hardware, and the localized ecological and health consequences so listeners grasp the full environmental cost of modern AI. I will also suggest simple practices and mindset shifts to reduce harm and encourage more cautious use of compute intensive systems. Attendees will leave better informed about AI's full environmental footprint, more cautious about deploying compute intensive systems, and equipped with simple practices to reduce harm.