Generative AI's Environmental Impact: Energy, Water, and Sustainability

MIT researchers have explored the significant environmental impact of generative AI, highlighting its substantial electricity and water consumption, especially in data centers. The study emphasizes the need for a comprehensive understanding of the environmental and societal costs of generative AI, considering the energy used for training and deploying models like GPT-3, and the environmental impact of manufacturing and transporting high-performance computing hardware.