**
As artificial intelligence rapidly evolves and permeates our daily lives, the environmental implications of this technology are coming under scrutiny. A growing concern is the significant water consumption associated with powering AI systems, particularly data centres. With various estimates circulating, it is essential to dissect the reality of AI’s water usage and its broader implications for sustainability.
Understanding AI’s Water Footprint
The rise of AI has led to an unprecedented demand for computational power, primarily managed through extensive data centres. These facilities are critical for running AI models like OpenAI’s ChatGPT and Google’s Gemini. However, they also have a substantial environmental footprint, particularly in terms of water use.
Data centres require cooling systems to prevent overheating, necessitating vast amounts of water. While estimates vary widely, research from the University of California suggests that ChatGPT consumes around 500ml of water for every 10 to 50 average responses. In contrast, Sam Altman, the CEO of OpenAI, claims that a single interaction with ChatGPT uses less than 1/15 of a teaspoon of water. Such discrepancies highlight the need for clarity on the true water costs of AI operations.
The Scale of the Problem
The magnitude of AI’s water consumption is alarming. A report from the UK Government Digital Sustainability Alliance indicates that by 2027, global water usage due to AI could escalate from 1.1 billion to a staggering 6.6 billion cubic metres. This figure equates to over half of the total water consumption in the UK. Given that nearly 68 per cent of data centres are located near protected biodiversity areas, the implications for local ecosystems and communities could be severe.
Critics argue that the escalating demand for water from data centres is exacerbating existing water scarcity issues. As AI continues to expand, the pressure on local water supplies intensifies, particularly in regions already facing challenges in accessing clean water. Experts warn that by the end of the decade, demand for fresh water could surpass supply by 40 per cent, with many data centres situated in high-risk river basins.
Perspectives on Water Use
Opinions on the severity of AI’s water consumption vary among experts. Some, like Andy Masley from Effective Altruism DC, assert that the water footprint of AI is negligible compared to other everyday activities. He points out that generating hundreds of thousands of prompts would require less water than producing a single pair of jeans, which consumes around 7,500 litres. Masley contends that the majority of our water usage occurs off-site and is largely invisible to consumers.
Conversely, researchers like Sam Gilbert from the University of Cambridge stress that the real issue lies not in individual usage but in the cumulative impact of data centres on local water resources. He advocates for greater transparency regarding the environmental footprint of these facilities, urging companies to disclose their annual water consumption.
The Future of AI and Water Sustainability
As the debate around AI and water consumption continues, it becomes evident that the future of digital technology must consider sustainability. Shaolei Ren, an engineering professor at the University of California, highlights the seasonal nature of water resources, suggesting that many data centres peak during summer months, exacerbating pressure on local water systems. Thames Water has even warned that data centres could face restrictions on their water usage during extreme heat events.
Interestingly, AI may also offer solutions to water conservation challenges. It has already been employed to enhance leak detection and improve water distribution efficiency. For instance, a water company in Surrey has successfully implemented AI technology to minimise leaks across its infrastructure. The World Economic Forum has reported that AI-enabled solutions in the U.S. could reduce water use by up to 15 per cent.
Why it Matters
The intersection of artificial intelligence and environmental sustainability raises critical questions about the future of technology in a resource-constrained world. As AI continues to shape our lives, understanding its hidden costs—particularly in terms of water consumption—is paramount. The challenge lies not just in minimising the water footprint of AI but also in ensuring that technological advancements do not compromise the health of our ecosystems. As we march towards a more digitised future, a commitment to sustainable practices must be at the forefront of innovation, paving the way for a balance between technological growth and environmental stewardship.
