Artificial intelligence drives increased energy and water consumption by data centers
Artificial intelligence and the data centers it requires are significant consumers of energy and water, contributing to climate change and resource depletion. Experts advocate for reduced AI use by individuals and greater transparency from tech companies on their environmental impact. Data center development is also facing growing public opposition in populated areas due to these concerns.
The Associated Press reports from Washington that artificial intelligence (AI) and the data centers it necessitates are major contributors to energy and water consumption, exacerbating climate change and resource scarcity. Experts, including Sasha Luccioni of the Sustainable AI Group and Kaveh Madani from the United Nations University Institute for Water, Environment and Health, highlight that AI's demands run counter to decarbonization efforts and criticize AI companies for their lack of transparency regarding resource usage.
Arecent report from the United Nations University indicates that global data centers consumed 448 trillion watt-hours of electricity last year, with projections to more than double in four years. By 2030, the water required for data center cooling could reach nearly 2.5 trillion gallons annually. Experts advise individuals to minimize AI use for simple tasks and to keep queries concise to lessen computational load. They also express concern over Big Tech's pervasive integration of generative AI into everyday tools, which limits user choice.
Ana Pinheiro Privette, formerly a sustainability official at Amazon Web Services, and University of Michigan computer science professor Mosharaf Chowdhury, underscore the absence of transparent data from AI companies on their environmental footprint, hindering informed consumer decisions. They suggest that strong consumer demand for sustainability could compel companies to disclose more. Balaji Tammabattula, COO of BaRupOn, acknowledges the escalating public "backlash" against data center construction in densely populated regions, referencing the 2.1 billion gallons of water used by data centers in two Virginia counties in 2023. Tammabattula emphasizes the need for data center developers to collaborate with communities to address environmental concerns, asserting that while AI's growth is inevitable, it must proceed sustainably.