Is A.I. really the monster it’s said to be?
By Riley Duignan
A.I., it’s the talk of the town, it’s a student’s best friend and middle aged person’s new google search bar. Everyone is bound to hear the debate of how much impact A.I. (especially generative) really on the earth, but are these people correct? Is your use of artificial intelligence really as harmful as it is helpful?
Artificial intelligence is a concept that’s been developing even since the 1960s, with the progression of technology also progressing A.I. ‘s advancement. Generative AI, the type of artificial intelligence used by the average person, beginning to take off in 2022, with ChatGPT being the most used A.I. tool up to now. In addition, over 34% of people in America are using ChatGPT. Generative artificial intelligence and society go hand in hand; as generative A.I. learns patterns and advances from human patterns, it generates content better and better. This content can start from analyzing your school homework to generating a picture of you and your grandmother, who passed away 5 years ago. Since the take off of A.I, other large technology companies have made their own A.I. tools or completely separate A.I. media. This includes Google, Amazon, Microsoft, Duolingo, Meta, Apple, and other large companies. But so many companies are doing this, but where and how is this A.I. being run? It’s beyond just your phone screen, and it’s actually everywhere.
A.I. is run through different data centers and servers that build off of previous knowledge, patterns, images, etc. This data is processed and eventually ready for questions and quick generation. Data is held in large computers and hardware, called GPUs. Due to the 24/7 running, these GPUs are bound to overheat, so for that reason, people have come up with many ways to cool down these large data centers. Water and evaporation cooling is one of the most common cooling methods; as water evaporates, it assists and regulates the temperatures of these GPUs. According to EESI, one 100-word A.I. prompt uses around 1 bottle of water, and a single data center uses up to 5 million gallons of water per day for this temperature regulation. Closely related, it is estimated that in 2027 A.I. data centers will use up around 1.7 trillion gallons of water annually. Communities around data centers in areas like Oregon, Arizona, and Indiana have sued large A.I. companies like Google and Amazon for causing water scarcity in the area.
Due to the process of evaporation, around 80% of this water is evaporated and is non-reusable. It is not a recycled resource, which is actively stripping water from already scarce areas.
Foodandwaterwatch.org warns, “As of 2023, nearly 80% of the water consumption of Google A.I. data centers in the U.S. came from drinking water sources.” Closely, in Arizona, data centers are estimated to use up over 177 million gallons of water per day, and 26 of them were placed in areas of initial water stress. This is causing farmers’ fields to experience excessive scarcity and dryness, while people neighboring or somewhat near these data centers have noticed an actual decrease in water pressure coming into their houses. A business insider article quotes, “The largest data centers can guzzle millions of gallons of fresh water a day. Roughly 40% of U.S. ones, Business Insider found, are in the most water-stressed areas of the country.” And in China, A.I. data centers use around 1.3 billion cubic meters of water [around 343.42 billion gallons] annually, and this is projected to more than double to 3 billion cubic meters in 2030. Cwrr.org states, “48% of China’s national data centre racks are located in the Dry 10, which are as dry as the Middle East. – At least 41% of China’s national data centre racks are in areas that are highly prone to drought, while at least 28% are in areas that are highly prone to floods; at least a fifth are very prone to both.” This would logically affect around 26 million people and their access to clean water or any water at all.
If this doesn’t alert someone, I don’t know what will. But there are things we and you can do as a collective.
Begin using Ecosia as a search engine.
(Exactly like Google), But, the more you search and the more ads are promoted through each search, a tree is planted. Ecosia partners with multiple non-profit organizations to plant trees. Ecosia makes it’s money through consumers clicking on search ads.
Use A.I. scarcely
Occasional A.I. use is okay, but consistent daily A.I. use only increases your reliability and decreases its sustainability. This may not be as sustainable as you think. You’re not the only one who says, “It’s okay if I use it occasionally.”, Imagine millions of other people also doing it “occasionally.” A.I. will still get just as much usage with this mindset, so be mindful.
Green A.I. platforms
Use A.I. platforms or search engines that promise sustainability, low water usage, or something in return. This includes planting of trees, decreased water usage, delivering water to scarce areas, etc.
Whether you think A.I. is a “monster” is solely your opinion, but the undeniable impact it is having on communities and the environment itself is argumentative. But knowing this, A.I. may just be a classic tale of a wolf in sheep’s clothing.
Sources
https://www.eesi.org/articles/view/data-centers-and-water-consumption
https://www.fastcompany.com/91454609/ais-future-runs-on-water
https://iee.psu.edu/news/blog/why-ai-uses-so-much-energy-and-what-we-can-do-about-it
https://www.bloomberg.com/graphics/2025-ai-impacts-data-centers-water-data/
CWR-2024-China-ICT-running-dry-The-rise-of-AI-climate-risks-amplify-existing-water-risks-faced-by-thirsty-data-centres.pdf
https://www.scmp.com/news/china/science/article/3259230/chinas-growing-data-centres-and-ai-industry-could-strain-scarce-water-resources-according-new-report
What the AI Data Center Boom Means For Our Water and Climate
https://www.businessinsider.com/how-data-centers-are-deepening-the-water-crisis-2025-6#:~:text=How%20Data%20Centers%20Are%20Deepening%20the%20Water%20Crisis%20%2D%20Business%20Insider&text=Data%20centers%20are%20critical%20to,resources%20in%20the%20American%20West.&text=The%20largest%20data%20centers%20can,stressed%20areas%20of%20the%20country.&text=You’re%20currently%20following%20this,the%20link%20in%20your%20email.
https://spectrum.ieee.org/ai-water-usage#:~:text=But%20in%20some%20regions%20where,in%20high%20water%2Dstress%20areas.