**
As the technology behind artificial intelligence (AI) continues to evolve at a breakneck pace, the environmental implications of its consumption patterns are coming under scrutiny. One pressing question is how much water is required to support the operation of AI systems like ChatGPT and Google Gemini. While some estimates suggest a minimal impact, others highlight alarming potential increases in global water usage that could exacerbate existing scarcity issues.
Understanding AI’s Water Footprint
The energy demands of AI are significant, and these demands extend to water usage, primarily for cooling the data centres that host the software. Each user prompt necessitates a certain amount of water to keep systems from overheating—a process that has sparked a debate about the overall environmental cost of AI.
For instance, Sam Altman, CEO of OpenAI, asserts that a typical query to ChatGPT consumes less than 1/15 of a teaspoon of water. In contrast, a study conducted by Google Gemini estimates an average prompt uses only 0.3 millilitres. However, these figures stand in stark contrast to research from the University of California, which suggests that ChatGPT may consume around 500 millilitres of water for every 10 to 50 medium-length responses. Moreover, a report from the UK Government Digital Sustainability Alliance warns that AI’s water consumption could soar from 1.1 billion to 6.6 billion cubic metres by 2027—an increase that could surpass half of the UK’s total annual water usage.
The Role of Data Centres in Water Consumption
Data centres, the backbone of AI operations, rely heavily on water for various functions. Cooling systems, electricity generation, and hardware manufacturing all contribute to their water footprint. According to the Lincoln Institute of Conventional Policy, a mid-sized data centre can use as much water as a small town, while larger facilities may consume up to five million gallons daily—equivalent to the needs of a city with a population of 50,000.
This burgeoning demand for water raises alarms, particularly in regions already facing water scarcity. The Government Digital Sustainability Alliance highlights that nearly 68 per cent of data centres are located near vital biodiversity areas, where ecosystems and local communities depend on clean water supplies. As AI technology expands, the pressure on these resources is expected to intensify, with predictions indicating that fresh water demand could outstrip supply by 40 per cent by the end of the decade.
Diverging Views on the Water Crisis
The debate surrounding AI’s water consumption has elicited varied opinions among experts. While some argue that the overall impact is minimal compared to traditional water usage, others caution against underestimating the potential crisis. Andy Masley, director of Effective Altruism DC, claims that the water footprint of AI interactions is relatively small, suggesting that the volume of water used by numerous ChatGPT prompts pales in comparison to the 7,500 litres required to produce a single pair of jeans, as noted by the UN.
Conversely, Nick Couldry, a sociologist at the London School of Economics, underscores the need to consider the broader implications of increased data processing reliant on AI technology. He warns that even if individual water usage remains low, the collective demand driven by an AI-dependent economy could lead to unsustainable environmental pressures.
Shaolei Ren, an engineering professor at the University of California, echoes these concerns, emphasising the seasonal and local nature of water resources. He points out that many data centres experience peak water usage during summer months, straining public water systems. Thames Water has issued warnings that data centres may face restrictions during extreme heat, highlighting the urgent need for a balanced approach to water management.
Can AI Help Mitigate Water Issues?
Despite the concerns surrounding water consumption, AI also holds promise for enhancing water efficiency in other sectors. Innovative technologies powered by AI can assist in leak detection and improve water distribution systems. For example, a water company in Surrey has begun employing AI to minimise leaks, and the World Economic Forum has reported that fully implemented AI solutions in the United States could reduce water usage by as much as 15 per cent.
Notably, Google has positioned its data centres as among the most efficient globally. A spokesperson for the tech giant stated that its facilities utilise air-cooling methods to minimise water use and commit to replenishing 120 per cent of their water consumption on average, thereby supporting local watershed health.
Why it Matters
The debate over AI’s water consumption highlights a critical intersection between technological advancement and environmental sustainability. As the demand for AI solutions continues to grow, so too does the imperative for transparency and accountability from tech companies regarding their resource consumption. Balancing innovation with environmental stewardship will be essential in ensuring that the benefits of AI do not come at the expense of our planet’s most vital resources. The conversation around AI and water is not just about numbers; it is about the future of our ecosystems and communities in a rapidly changing world.