AI, data centres, and their impact on the environment

Over the past couple of years, different AI generation tools such as OpenAI’s ChatGPT-4 have increasingly been adopted for both professional and recreational use. More workplaces and sectors have incorporated AI tools into their work processes, creating demand for further investment into Artificial Intelligence. With this growing demand, there is also growing concern over the resulting environmental impacts. The strain on worldwide water supplies is of particular concern, given the vast quantities of water required to maintain the data centres responsible for AI content.

While tools such as chatbots have existed since the 1960s, this concept has developed rapidly across the 2020s to the stage where AI has become an integral part fas to how many people research, work or create content to be put online, despite the concerns brought forward by environmentalists. 

The energy required and consumed by data centres have skyrocketed in recent years due to public use of AI tools, with some of these data centres matching the energy consumption of whole countries like Morocco, according to a review of permits for data centres by Wired. It is predicted that they will be in competition with some of the biggest energy generators on the Earth, such as Japan, by 2030, according to a study by the International Energy Agency.

Many different technological and environmental experts have spoken for and against AI usage. In a recent speech at the University of Arizona, Google CEO Eric Schmidt said that AI was “larger, faster, and more consequential” than the tech that has come before, resulting in a negative reception from the crowd of graduating students. Different experts have spoken out against AI, such as Professor Elsa Olivetti, who said: “There are much broader consequences that go out to a system level and persist based on actions that we take”.

AI could result in a further contribution to already existing crises, particularly in regions where people are dealing with deadly temperatures, water shortages, or both. 

This impact on the environment and water supply has and could have further repercussions for countries that we support here at Concern. People living in extreme poverty or those already facing humanitarian crises could struggle to access clean water, resulting in health concerns such as dehydration and sanitation.

In particular, the spread of diseases like cholera is one of the biggest issues when people have a lack of clean water, due to people having to cook with and drink contaminated water.

Meanwhile, in places including Malawi, Somalia and northern Kenya, drought worsened by the impacts of climate change leads to a shortage of water for drinking, cooking and agricultural use - in turn risking failed harvests and a lack of food. This perpetuates a cycle of malnutrition that can be hard for vulnerable communities to escape. 

While almost every country in the world contains a data centre that is capable of producing AI content, the current leader of data centres as of 2026 is the US, with over 4,000 data centres.

Each one of these data centres is currently producing energy numbers in the thousands, with the annual energy output caused by data centres being approximately 415 terawatts per hour(TWh) annually, or over a billion kilowatts per hour(KWh). With the US leading the quantity of data centres worldwide, much of this energy consumption links back to many of the data centres in the US. This has resulted in higher carbon emissions from these data centres as many have resorted to using fossil fuels in order to power their generators. 

And of course, the chilled water that is used to cool down the generators is placing a massive toll on water supplies. MIT Climate worker Noman Bashir estimates that for each kilowatt per hour used by a data centre, two litres of water would be required to cool them down. By this estimate, data centres around the world are producing 415 TWh, or 415 billion kilowatts, using up billions of litres of water according to the International Energy Agency.

As AI fuels demand for increasing numbers of data centres, they currently make up 1.5% of the earth’s electricity consumption. Advanced AI generative systems are on their way to becoming a leading source of carbon emissions. It is estimated that a single use of one of the most popular and public AI tools, ChatGPT-4, creates roughly 4.32 grams of CO2. However, as we are still in the early stages of adopting these tools, they do not yet compare with other sectors' impact on the environment, such as the fuel industry and fast fashion.

With increasing development of AI generation, worldwide water supply shortages will hit the least responsible the hardest. Concern to put further focus onto countries dealing with water crises. Concern works to support many countries affected by the global water crisis to ensure that people have access to clean water to drink, clean, cook with, and use for irrigation.

Concern leads over 20 WASH (water, sanitation and health) programmes, to support impoverished countries. These programmes improve access to the supply and foundation required for people to get clean water. Concern’s goals of strengthening water systems has resulted in the establishment of different sustainable water sources, such as the building of water pumps in areas dealing with clean water shortages. This allows for people dealing with water crises to gain stronger nutritional, health and agricultural ability.

Access to safe, clean water is a human right, and is vital to prevent illness, disease and malnutrition, to grow crops, and to support a healthy lifestyle. While technology opens doors for us all, it must be sustainable, and we must continue to prioritise the needs of the vulnerable in the face of a growing crisis.