The increasing reliance of AI on vast amounts of water for its operation has brought Big Tech’s water footprint into the spotlight. Companies like Microsoft and Alphabet (Google’s parent company) have reported significant increases in water usage, primarily driven by advancements in AI technologies.
A recent study by Shaolei Ren at the University of California, Riverside, highlights this issue, revealing that OpenAI’s ChatGPT consumes about 500 millilitres of water for every 10 to 50 prompts.
“In general, the public is getting more knowledgeable and aware of the water issue and if they learn that the Big Tech’s are taking away their water resources and they are not getting enough water, nobody will like it,” Ren told CNBC via videoconference before adding that he thinks we are going to see more clashes over the water usage in the years to come, making this type of risk the responsibilty of the companies.
This consumption is notably high considering the hundreds of millions of monthly users interacting with AI models like ChatGPT. The study focuses on the potential environmental impact of AI’s water footprint, which could hinder the sustainable and responsible development of AI technologies. Despite requests for comments, OpenAI did not respond to the findings.
Data centres, crucial for Big Tech operations, require significant water resources to maintain server efficiency. Meta reports that its data centres are the primary consumers of both its water and energy resources, contributing heavily to its greenhouse gas emissions. Microsoft’s water usage rose by over a third between 2021 and 2022, totalling nearly 1.7 billion gallons — equivalent to over 2,500 Olympic-sized swimming pools. Similarly, Google’s water consumption in 2022 was 5.6 billion gallons, a 21% increase from the previous year.
Both Microsoft and Google are striving to become “water positive” by the end of the decade, aiming to replenish more water than they consume. However, the recent launch of their respective ChatGPT competitors, Bing Chat and Google Bard, could further escalate water usage.
Microsoft is actively investing in research to understand and mitigate the environmental impact of AI, including energy, water use, and carbon emissions. They are committed to becoming carbon-negative, water-positive, and zero waste by 2030. On the other hand, Google claims that while AI’s computing demand has surged, the energy required is rising slower than anticipated. The company is focused on reducing the carbon footprint of AI operations, asserting that their data centers deliver five times more computing power with the same amount of electricity compared to five years ago. Their latest TPU v4 supercomputer is touted as one of the most efficient and sustainable machine learning infrastructure hubs.
Featured image: Credit: Наталья Сотникова, PlaygroundAI
