**
As artificial intelligence continues its rapid evolution, a pressing question emerges: what is the environmental cost of our increasing reliance on AI technologies? Specifically, growing scrutiny is being placed on the significant water consumption associated with data centres that underpin these systems. This article delves into the complexities of AI’s water usage, the implications for sustainability, and the varying perspectives on the urgency of this issue.
The Thirst of Data Centres
At the heart of AI’s operations lies the data centre, a facility that houses the servers necessary for processing vast amounts of information. These centres require substantial amounts of water for cooling systems to prevent overheating, generating electricity, and during the manufacturing of the hardware itself. Reports indicate that a data centre of moderate size can consume as much water daily as a small town, while larger facilities may require up to 5 million gallons, equivalent to the needs of a city with a population of 50,000.
The water usage associated with AI technologies has sparked a debate about sustainability. Estimates vary widely, with Sam Altman, CEO of OpenAI, asserting that ChatGPT consumes less than 0.1 millilitres of water per query. In contrast, a study from the University of California in 2023 suggested that for every 10 to 50 medium-length responses generated, ChatGPT could use approximately 500 millilitres of water. Even more concerning, the UK Government Digital Sustainability Alliance anticipates that AI could increase global water usage dramatically, from 1.1 billion to 6.6 billion cubic metres by 2027.
Rising Concerns Over Water Scarcity
The implications of AI’s water consumption are particularly alarming in regions already facing water shortages. Critics argue that the increasing demand for water from data centres is exacerbating existing pressures on local supplies. A report from the Government Digital Sustainability Alliance highlights that roughly 68 per cent of data centres are located near areas of significant biodiversity, where ecosystems depend on clean water. As demand grows, the risk of water scarcity intensifies, with forecasts suggesting that freshwater demand could surpass supply by 40 per cent by the end of the decade.
The findings raise critical questions about the sustainability of our digital infrastructure. As AI usage expands, particularly in water-stressed regions, local communities may find their water resources increasingly strained. The situation is further complicated by the reality that many data centres are located in river basins at high risk of pollution, which threatens the safety of available water supplies.
Differing Perspectives on Water Usage
While many experts deem AI’s water consumption a serious global issue, others argue that the concerns may be overstated. Andy Masley, director of Effective Altruism DC, contends that the water footprint associated with AI is minuscule compared to other everyday activities. He notes that hundreds of thousands of queries to ChatGPT would require less water than the production of a single pair of jeans, which the UN estimates consumes around 7,500 litres.
Masley suggests that a user would need to generate over 1,000 prompts in a single day to marginally increase their daily water footprint. This perspective posits that, in some cases, remaining at home and engaging with AI could result in a lower water footprint than typical daily activities. However, Sam Gilbert from the University of Cambridge stresses that the concern should not be about individual usage but rather the cumulative impact of data centres on local environments and water supplies.
The Future of AI and Water Sustainability
As the conversation around AI’s water consumption continues, the potential for AI to assist in water conservation efforts is also being explored. Professor Shaolei Ren from the University of California notes that AI technologies can help detect leaks and improve the efficiency of water distribution systems. For instance, a water company in Surrey has begun utilising AI to reduce leaks across its network, demonstrating the dual role of AI as both a consumer and a potential saver of water resources.
Companies like Google are also taking steps to address these concerns. Their data centre in Waltham Cross employs air-cooling systems designed to minimise water usage. Additionally, Google claims to replenish 120 per cent of the water it consumes across its operations, showcasing an effort to mitigate its environmental impact.
Why it Matters
The intersection of AI technology and water sustainability is increasingly critical as we navigate the future of our digital landscape. With the potential for AI to both exacerbate and alleviate water scarcity issues, it is essential to strike a balance between innovation and environmental stewardship. As the demand for AI capabilities grows, so too must our commitment to understanding and addressing the ecological implications of our technological advancements. The decisions made today will significantly impact tomorrow’s water resources, necessitating a strategic approach that prioritises sustainability in the heart of Silicon Valley and beyond.
