Why AI is Ruining the Earth

So, we all know that AI (Artificial Intelligence) is new, hip, the new BIG thing companies want to get their grubby hands on. From Google, Facebook, X (formerly Twitter) we have seen AI pop up everywhere. Even in the most unlikely of places. At my usual walk-in clinic, they have posters up claiming they “use AI to ensure the doctors here can more efficiently help patients like you out and understand what your issue is.” You can turn left, right, upside down, and AI is everywhere. And if you uncover a corner where the chatbots haven’t infected yet, it’s only a matter of time.

Regardless of if you are a fan of AI or not, you have the right and responsibility to understand the impact of using sites like ChatGPT or Character.AI. No worries, I have the information right here, with sources, of course. I will do my best to explain how, and more importantly, WHY you should care about how AI affects the environment and why we should care.

AI companies run their servers through data centers. Data centers are heavy users of electricity and often, very often, use chilled water to cool down their systems. These are the “Cloud” your phone backs up your photos too. Though making sure you don’t lose your grocery list from April 2022 isn’t their only function. It runs and trains AI models. Training alone uses thousands of gigawatts of electricity and water. Not to mention the carbon gas emissions it creates. For example, Google owns and operates over 130 globally. President Donald Trump during his second inauguration claimed he only wanted to build more. Even though, as of 2025, we have over 11 800 data centers globally.

In 2024, the Biden Administration commissioned a report on the energy uses of these buildings. Reportedly, by 2028, the U.S. centers alone could consume more than 12% of the nation’s energy. Another study a year prior reported that servers' use of electricity increased from 2688 megawatts to 5341 in just one year. And consumed 460 terawatt-hours by the end of 2022. I know that’s a lot of numbers but lets put this into perspective. For training alone, it is estimated it takes roughly the same amount of power for 120 U.S. citizens homes and creates 552 tons of carbon dioxide. Now, I’m Canadian but I do know many Canadian and American citizens who have experienced extended lengths of time without power in their homes. Currently in Puerto Rico, it is a common occurrence that power will just randomly go out for days.

Water is an even bigger concern. But first a fun fact! Did you know that only 3% of the world’s water is fresh? Or that only 0.5% of that fresh water isn’t locked away in glaciers? Fun right? How about the fact that data centers use clean, fresh water to cool their servers?

Google is said to have used over 550 000 gallons of water in 2024. An average data center used only 18 000 gallons of water. Which is no sneeze but significantly less than Google. Companies have also started building data centers in places where in general water is already a limited resource such as Saharan Africa. AI is on the track to drink down over 4.2 billion cubic meters of water annually before 2030. Another way of looking at it is, a single kilowatt hour of energy use also is 2 liters of water usage. The use of AI pushes forward far past the safe line of water consumption. Once the thirsty robot has used the fresh water, it’s gone. Evaporated into gas. There is no way of getting that water back.

Though there are companies (Microsoft, Google, Meta) who are looking into ways of non-water consuming AI data centers, the damage it has already done is quite significant. "I think before generative A.I. came along in late 2022, there was hope among these data center operators that they could go to net zero. I don't see how you can, under current infrastructure investment plans, you could possibly achieve those net zero goals," Professor Ben Lee of the University of Pennsylvania told NPR in May 2025. Furthermore, Trump’s promise of data centers, he proudly calls “Stargate”, would also outweigh and further the damage AI servers do to the planet. Stargate would add another 15 gigawatts of power used. From what I could find, there has been no talk about slowing down energy use or minimizing carbon dioxide emissions.

So, big question, why should you care? If sucking the Earth dry of water like a vampire and blowing carbon gases into our faces like a lifted diesel truck isn’t bad enough. That’s water you could be drinking, energy you could be using, clean air you could be breathing for ages and ages. The introduction of AI has sped up the process of the world crumbling around us. 2 million U.S. citizens experience lack of clean drinking water or basic indoor plumping. The average monthly power bill in the same country is $150 dollars in the summer and $200 in the winter. It is insane to have to spend that much money on something that is a necessity. Something I believe is one of the many things we shouldn’t have to pay for in the first place. (An article on that at a later date.)

There are suggestions backed by scientists on how we can slow down AI consumption of our resources. Transparency is a big one. Senator Edward Markey (Massachusetts) has set into motion a bill that would make it so AI companies would have to be open about how much resources it is using. Many other senators are following suit; it’s a matter if their bill gets passed or not. Technological advances, such as data centers or chips that do not need a coolant, would also help tremendously. But the biggest thing is making sure that they, the government and companies, know about climate change and how AI is only speeding up the process.

Please remember this the next time you open Dall E. for an image of a monkey impersonating Elvis. A few generated pictures could’ve served drinking water to a large portion of the population and powered a lot of homes. Those pictures choke out real monkeys who breathe in those carbon gases. Those pictures, over-time, might be the final ones the Earth ever sees.

I’d like to thank:

NPR: https://www.npr.org/2025/05/07/1249592906/energy-water-ai-climate-tech

MIT:https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117

The University of Illinois: https://cee.illinois.edu/news/AIs-Challenging-Waters

For their papers and research into the dangers of AI environmentally. If you’d like to learn more I highly recommend giving these articles/papers a read on your own time. I’d also like to thank you for reading this article. Please share this with a friend to spread the word of how to heal our sick Earth. We only have one.

Melody Vanbeak