**
As artificial intelligence (AI) continues to evolve at a breakneck pace, a new concern has emerged regarding its environmental impact—specifically, the substantial amount of water required to support these technologies. With data centres at the heart of AI operations, the implications of their water usage raise pressing questions about sustainability and resource management.
The Water Footprint of AI
AI platforms, such as ChatGPT and Google Gemini, rely on extensive data centres that consume considerable amounts of water for cooling systems and hardware production. The estimates surrounding AI’s water consumption vary widely. Sam Altman, CEO of OpenAI, has claimed that each interaction with ChatGPT uses less than 1/15 of a teaspoon of water. In contrast, a study conducted by Google suggests an average query consumes less than 0.3 millilitres. However, research from the University of California in 2023 indicates that ChatGPT may actually require approximately 500 millilitres of water for every 10 to 50 medium-length responses.
This discrepancy in figures illustrates the ongoing debate about the real environmental cost of AI technology. Alarmingly, a report from the UK Government Digital Sustainability Alliance predicts that AI could push global water usage from 1.1 billion to 6.6 billion cubic metres by 2027—equivalent to over half of the UK’s total water consumption.
Understanding the Demand
Data centres are not merely hubs of digital activity; they are also significant consumers of local water resources. These facilities require water to cool their systems, generate electricity, and manufacture the hardware they depend on. According to the Lincoln Institute of Conventional Policy, a mid-sized data centre can use as much water as a small town, while larger centres might draw up to 5 million gallons per day, comparable to the needs of a city of 50,000 inhabitants.
As these facilities proliferate, concerns are mounting. Local communities, particularly in water-scarce regions, worry that data centres will exacerbate existing water shortages. The Government Digital Sustainability Alliance has highlighted that nearly 68 per cent of data centres are located near protected biodiversity areas, which further intensifies the risk of water scarcity for ecosystems that rely on clean water.
The Controversy Surrounding Consumption Estimates
Experts are divided on whether AI’s water consumption constitutes a genuine crisis. Some argue that the water used per individual interaction is minuscule when placed alongside other daily water expenditures. For instance, Andy Masley of Effective Altruism DC points out that the water footprint of hundreds of thousands of prompts is still less than the 7,500 litres required to produce a single pair of jeans.
However, the crux of the issue lies not in individual usage but rather in the cumulative impact of numerous data centres on local water supplies. Sam Gilbert from the University of Cambridge emphasises the need for data centre operators to be transparent about their environmental footprint, as the potential impact on surrounding communities could be profound.
Conversely, sociologist Nick Couldry warns that regardless of the accuracy of water usage estimates, the overall increase in data processing demands driven by AI will inevitably challenge environmental sustainability. The addictive nature of AI technologies could lead to unsustainable pressures on local ecosystems and water resources.
Seasonal Pressures on Water Systems
Seasonal spikes in water demand pose a significant challenge for public water systems, particularly during summer months. Engineering professor Shaolei Ren notes that many data centres experience peak water usage during hot periods, straining local resources. Thames Water has previously cautioned that data centres may face restrictions during droughts, highlighting the urgent need for effective water management strategies.
Ren argues that while total water availability may seem sufficient, the localised nature of water resources complicates the situation. Expanding water infrastructure to meet peak demands is costly and often impractical, which necessitates a more nuanced understanding of water consumption in relation to location and timing.
AI’s Potential Role in Water Conservation
Interestingly, AI may also offer solutions to mitigate water waste in other sectors. For example, advanced technologies are being deployed to identify leaks and optimise water distribution networks. A water company in Surrey has begun utilising AI to reduce leaks effectively. The World Economic Forum has reported that the full implementation of AI-driven water solutions in the United States could lead to a 15 per cent reduction in water consumption.
In a bid to address these concerns, Google has claimed that its data centres are among the most efficient globally, employing air-cooling techniques to minimise water usage. Furthermore, the company asserts a commitment to replenishing 120 per cent of the water it consumes, signalling a proactive approach to environmental stewardship.
Why it Matters
The intersection of AI technology and water consumption raises critical questions about sustainable development in a rapidly digitising world. As the demand for AI continues to escalate, so too does the pressure on local water resources. Balancing the benefits of AI with ecological responsibility will require transparency, innovation, and a robust commitment to sustainable practices across the industry. The future of both technology and the environment hinges on our ability to address this urgent challenge.
