Should I worry about how much water my AI chatbot conversations are using?
A.It is stated that the thirst for water has led to widespread environmental fears, with many people concerned that rapidly advancing technology is putting more pressure on the world’s resources.
Every question or question a person feeds into the AI will require energy and water to cool the data center containing the software. Estimates of how much water AI uses are widely debated, with different AI companies reporting varying figures.
Sam Altman, OpenAI’s CEO, said ChatGPT uses less than 1/15 teaspoon for an average query. A Google Gemini study claims that the average AI prompt uses less than 0.3ml of water.
But other estimates suggest he used much more. A 2023 study at the University of California calculated that ChatGPT “drinks” approximately 500ml of water for every 10 to 50 medium-length responses.
A report by the UK Government Digital Sustainability Alliance predicts that AI could increase global water use from 1.1 billion to 6.6 billion cubic meters by 2027; This amount is equivalent to more than half of the UK’s total water use.
Why does artificial intelligence use water??
Data centers running software like ChatGPT or Google Gemini rely on water to cool systems and prevent overheating.
They also use water to produce electricity and the equipment they work on.
The Lincoln Traditional Policy Institute said a medium-sized data center would consume as much water as a small town, and a larger data center requiring up to 5 million gallons of water each day would consume as much water as a city of 50,000 people.
Why are people worried about this?
People are concerned about AI’s water consumption because they fear it will cause increased demand on already limited resources.
Data centers put pressure on surrounding communities by using local water resources; people fear this situation will worsen as AI expands, especially in regions where water is already scarce.
Members of the government’s Digital Sustainability Alliance Planetary Impact working group said almost 68 per cent of data centers were close to protected or important biodiversity areas, where ecosystems depend on clean water supplies and communities depend on them.
As the demand for water increases, water scarcity and water stress are becoming more common problems.
The group said: “Demand for fresh water is expected to exceed supply by 40 per cent by the end of the decade, and 55 per cent of global data centers are located in river basins at high risk of water pollution, meaning much of local water may be unsafe for use, increasing pressure on clean water supplies and worsening overall water scarcity in regions.”
Is this a problem?
Many experts say that the amount of water consumed by artificial intelligence is a global crisis. But others say the fears are exaggerated.
Andy Masley, director of Effective Altruism DC, a nonprofit focused on improving society through research, argues that the amount of water an individual uses is much less than most people assume.
He said hundreds of thousands of ChatGPT “requests” would require less water than your pair of jeans, which the UN says requires about 7,500 liters of water to produce.
“This is incredibly small by the standards most people use water in their daily lives,” he said. Independent. “Almost all of our water footprint is invisible to us because it occurs off-site and elsewhere.”
Mr. Masley estimates that a person would need to send out more than 1,000 prompts a day to increase their daily water footprint by just one percent.
Staying home and generating so much stimulation may actually result in a smaller water footprint than going out and using electricity, he said.
The water footprint measures the total amount of freshwater used to produce a product. The European Union now requires data centers to report annual freshwater consumption.
Sam Gilbert, a researcher at the Bennett School of Public Policy at the University of Cambridge, said the issue was not individual consumption but the impact of the centers on their immediate surroundings and the demand on local water supplies.
He said companies building and using these data centers need more transparency about what their true environmental footprint will be.
Mr Gilbert said the estimate that ChatGPT used 500ml of water for every 10 to 50 replies was “probably overstated”.
“But even if it were true, that’s not a lot of water considering the amount of water people use in their daily lives,” he said.
But sociologist Nick Couldry, of the London School of Economics, said: “Whatever the competing calculations for water use, we must consider the sustainability of the vastly increased data processing that an economy and society heavily reliant on AI will require.”
Even if water use could be reduced, he said, tech companies “want and need us to use AI continuously for a longer part of our lives.”
“It is difficult to understand why this addictive business model will not lead to unsustainable demands on the physical environment, and competing versions of AI development will require even more energy,” he added.
One of the main problems is that many data centers have high water use during the summer months, which puts great pressure on the public water system, said Shaolei Ren, an engineering professor at the University of California.
Thames Water has previously warned that data centers could face usage restrictions during the hottest and driest times of the year.
Mr Ren said: “Water is a local and seasonal resource. There is plenty of water in total, but not everywhere or every time we need it. Looking only at total volume without considering the context of location or timing can miss important nuances.”
“Water infrastructure needs to be sized to support peak demand. However, increasing water infrastructure capacity is extremely costly for public water systems.”
Can AI help save water in other processes?
Mr Ren said AI could also help save water from other processes. Some of the ways it is currently doing this is through technology that can detect leaks and improve energy-efficient water delivery.
In 2024, a water company in Surrey began using AI to reduce leaks on its network. The World Economic Forum reported that if AI-powered water solutions were fully implemented in the United States, they could reduce water use by 15 percent.
Google’s data center in Waltham Cross uses air cooling to limit the amount of water it uses. A spokesperson told Independent: “As a pioneer in computer infrastructure, Google’s data centers are some of the most efficient in the world.
“Beyond our operations, Google is committed to improving local watershed health and replacing an average of 120 percent of the water it consumes where it operates office campuses and data centers.”
