**
As artificial intelligence continues to dominate the technological landscape, an unexpected concern has emerged regarding its environmental impact—specifically, the significant amounts of water required to support the data centres powering these advanced systems. This issue has sparked a heated debate about the sustainability of AI technologies and their consumption of local water resources. With projections indicating a dramatic rise in water demand linked to AI, stakeholders must grapple with the implications of this trend.
The Water Consumption Debate
The conversation surrounding AI’s water usage is rife with conflicting estimates. For instance, Sam Altman, CEO of OpenAI, claims that ChatGPT consumes less than 1/15 teaspoon of water for each query. In contrast, a study conducted by Google Gemini suggests an average AI prompt uses no more than 0.3 millilitres of water. However, research from the University of California in 2023 presents a starkly different picture, estimating that ChatGPT requires approximately 500 millilitres of water for every 10 to 50 medium-length responses.
Adding to the complexity, a report by the UK Government Digital Sustainability Alliance warns that AI could escalate global water consumption from 1.1 billion to 6.6 billion cubic metres by 2027—an amount that exceeds half of the total water usage in the UK. As AI technologies proliferate, the strain on water resources could become increasingly unsustainable.
Why Water is Essential for AI Operations
Data centres, the backbone of AI systems like ChatGPT and Google Gemini, require substantial amounts of water primarily for cooling purposes. These facilities generate significant heat, and without adequate cooling, they risk overheating. In addition to cooling, water is also essential for electricity generation and the manufacturing of the hardware that supports these operations. The Lincoln Institute of Conventional Policy has reported that a mid-sized data centre consumes as much water as a small town, while a larger facility may utilise up to 5 million gallons daily—equivalent to the water needs of a city with a population of 50,000.
Growing Concerns Over Local Water Supplies
The environmental implications of AI’s water consumption are becoming increasingly alarming. Critics are particularly concerned about the pressure being placed on local water supplies, especially in regions already grappling with scarcity. According to members of the Planetary Impact working group within the Government Digital Sustainability Alliance, nearly 68% of data centres are situated near vital biodiversity areas, where local ecosystems depend heavily on clean water. The escalating demand for water could exacerbate existing shortages, with projections suggesting that fresh water demand might outstrip supply by 40% by the end of the decade.
This growing tension between AI demand and local water resources raises significant ethical questions. Communities near data centres may find their water access compromised as these facilities draw from local supplies, further contributing to the environmental strain.
The Counterargument: Is the Fear Overblown?
While many experts highlight the potential crisis stemming from AI’s water consumption, some argue that the fears may be overstated. Andy Masley, director of Effective Altruism DC, contends that the individual water footprint associated with AI usage is minimal when compared to other daily activities. He notes that hundreds of thousands of queries to ChatGPT would consume less water than the production of a single pair of jeans, which requires around 7,500 litres.
Masley points out that a person would need to submit over 1,000 prompts in a day to increase their daily water footprint by just 1%. Thus, for the average user, the water demand from AI interactions may be negligible compared to other common water-consuming activities.
The Need for Transparency and Accountability
Despite the varying perspectives on water usage, there is a consensus that greater transparency is essential. Sam Gilbert, a researcher at the University of Cambridge, argues that the focus should not solely be on individual consumption but rather on the broader environmental impact of data centres and their demands on local water supplies. He stresses the necessity for companies to disclose the environmental footprints of their data operations, thereby fostering accountability.
Additionally, Nick Couldry, a sociologist at the London School of Economics, warns that regardless of water usage calculations, the burgeoning demand for AI will inevitably pose sustainability challenges. He believes that the addictive nature of AI technologies will lead to escalating pressure on natural resources, necessitating a critical examination of our reliance on these systems.
Innovative Solutions: Can AI Help Itself?
Interestingly, AI may also offer solutions to its own water consumption challenges. Professor Shaolei Ren from the University of California highlights the potential of AI technologies to optimise water use in other sectors. For example, AI has been deployed to detect leaks and improve the efficiency of water distribution systems, leading to significant reductions in water waste.
In Surrey, a water company has begun utilising AI to address leaks in its network. Moreover, Google claims to be at the forefront of water-efficient data centre operations, employing air-cooling systems to minimise water use and pledging to replenish 120% of the water consumed in its operations.
Why it Matters
The implications of AI’s water consumption extend far beyond the tech industry, touching on critical issues of sustainability, resource management, and social equity. As AI continues to evolve and integrate into our daily lives, the environmental costs associated with its infrastructure must not be overlooked. A proactive approach is essential, fostering transparency and accountability within tech companies while exploring innovative solutions to mitigate resource strain. The conversation around AI and water usage is not merely a technical concern but a pressing societal issue that demands our attention and action.
