I’ve heard it, you’ve heard it: AI is terrible for the environment. But why?

Well, here’s what Google AI has to say about it, funnily enough:

In all seriousness, this is slowly becoming a huge problem. Especially when it comes to things like Meta and Google AI that we cannot opt out of (if you know how, please let me know, I’m a Boomer and cannot figure it out). We have no choice but to partake in using AI every day.

And with AI’s extreme rise in the last few years and with just how new it is for us, it’s hard to predict its true environmental impact. But, researchers do know that it is of concern. No one is saying not to use AI, but we need to tackle these problems now before it gets out of hand.

First, let’s talk about how AI might be a benefit to the environment. According to the UN Environment Programme (UNEP), “It can detect patterns in data, such as anomalies and similarities, and use historic knowledge to accurately predict future outcomes. That could make AI invaluable for monitoring the environment and helping governments, businesses, and individuals make more planet-friendly choices. It can also enhance efficiencies. Advances like those are fostering hope that AI could help the world tackle at least some aspects of the triple planetary crisis of climate change, nature and biodiversity loss, and pollution and waste.”

Another app is xView2, a program that combines machine-learning models and computer vision with satellite imagery to identify buildings damaged in natural disasters. This one is even more eerie as we learn more throughout this post. AI is partaking in destroying the planet and causing these natural disasters so we create AI to help people after these disasters. Dystopian.

Of course, this comes at a cost.

Building the data centers

Like other tech, these data centers require vast amounts of materials to be created, all of which are rare earth minerals and metals like cobalt, coltan, gold, and more. According to UNEP, in order to make a 2 kg computer it requires 800 kg of raw materials.

Not only is the extraction of these metals harmful to the environment but also to the people who are doing the extraction. They are likely in the global south, do not have protective equipment, are underage (as young as 10), are risking their health and lives, and are not paid a fair wage. Many of them are also kicked off their own lands in order to turn the land into mines. We learn more about that in this video, the truth about EVs.

On the other end of this production chain is e-waste. Old machines are thrown away or hopefully recycled, but there will always be harm in this process as well. Usually, this produces mercury and lead which get into our groundwater and soil. It’s an old video, but you can learn more about where our e-waste actually ends up and the environmental impact of e-waste as a broad category here.

It is important now while AI is new that recycling procedures are implemented. This should not be an afterthought, it should be the norm right off the bat to ensure as little e-waste as possible. Tech is extremely valuable to recycle. It not only reduces pollutants but could literally save lives.

Energy Usage

It not only takes vast amounts of energy to operate these data centers, but also to train the AI before it hits consumer use. Earth.org states that “according to OpenAI researchers, since 2012, the amount of computing power required to train cutting-edge AI models has doubled every 3.4 months. By 2040, it is expected that the emissions from the Information and Communications Technology (ICT) industry as a whole will reach 14% of the global emissions, with the majority of those emissions coming from the ICT infrastructure, particularly data centres and communication networks.”

Here’s a chart that helps put into perspective what that looks like. Again, this is just TRAINING the AI.

The International Energy Agency states that one search using ChatGPT consumes 10 times the electricity than that of one Google Search. This is especially problematic since most data centers still operate using fossil fuels. Of course, we should also conserve energy when using green energy, but this is even more of a problem since greenhouse gases are the leading cause of climate change and these data centers are burning high amounts of coal and oil. According to the UNEP “the International Energy Agency estimates that in the tech hub of Ireland, the rise of AI could see data centers account for nearly 35% of the country’s energy use by 2026. ” According to the Harvard Business Review (HBR), “In the United States, the rapidly growing AI demand is poised to drive data center energy consumption to about 6% of the nation’s total electricity usage in 2026.”

And, in just the last 13 years, the number of AI data centers has increased from around 500,000 in 2012 to over 8 million in 2024. Of course, this need will continue to go up. More data centers = more energy, more e-waste, and more water usage as we will explore next.

Note: it is hard to calculate precisely how much AI uses and its precise impact on climate change since there are so many different types of AI and therefore many different rates of energy consumption.

Water Usage

I had heard this being thrown around all over the internet in the last few months: AI uses a lot of water. But I never saw solid numbers until now. According to UNEP, “globally, AI-related infrastructure may soon consume six times more water than Denmark, a country of 6 million, according to one estimate.” Of course, as AI becomes more and more popular as the Earth.org predictions suggest it will, that means more water usage down the line. This is a huge problem. Already about 25% of the world struggles with access to clean drinking water. And with climate change-induced droughts on the rise, this number will also only continue to grow. AI needing more water and more humans running out of water is a bad combination we should tackle now and not later.

In the US, where the average per capita water withdrawal is 132 gallons per day, a large data center consumes water equivalent to that of 4200 people. This makes data centers one of the top 10 of “water-consuming industrial or commercial industries” in the country. As with energy, it’s hard to calculate precisely how much water each act of using AI uses since it all functions a bit differently. But, the easiest way to overcome this is to use it sparingly on our end and for the AI companies to build their facilities in areas where water is not scarce.

Overconsumption

AI is good…a little too good. Have you ever Google searched something and how your YouTube and Instagram ads are about that exact search you just did? Yeah, that’s the whole point. Brands will pay to advertise and they will get the most bang for their buck thanks to AI. Advertisers now don’t even have to try to get your attention, AI does all the work for them in order for you to buy more and more stuff.

This probably won’t affect those of us reading this as much as the average person since we’re already eco-conscious about our purchases, but this will most certainly lead to an uptick in consumer spending. According to Scientific American, “targeted ads push a steady rotation of cheap, mass-produced clothes to consumers, who buy the outfits only to replace them as soon as a new trend arrives. That creates a higher demand for fast-fashion companies, and already the fashion industry is collectively estimated to produce up to eight percent of global emissions. Fast fashion produces yet more emissions from shipping and causes more discarded clothes to pile up in landfills.”

Yet another way that fast fashion is contributing to killing the planet.

Who is this affecting?

A tale as old as time, the wealthy and well-off will not see much of these consequences. Though as AI becomes more and more popular, fewer people are out of harm’s way from its impact. If you live near a data center, that’s your fresh drinking water that will be evaporated to cool it down. I hope you don’t live in a drought area. AI requires an abundance of fuel. If you live near oil extraction sites, that’s your air, soil, and water that is going to be polluted.

Of course, this will primarily affect those in the global south: those who are mining for these resources and those already living in climate change-induced drought areas like Argentina and Chile according to HBR.

Think about how landfills and other toxic sites are chosen: they target people of color and low-income neighborhoods. I don’t think these data centers will be any different.

Can we use it sustainably and ethically?

Can we move forward in this age of AI while still caring about the planet? Yes!

Here are some things that we can do:

Limit your usage of AI. Don’t use it willy-nilly. Do a Google Search (or better yet, Ecosia since they plant trees while you search) instead of using ChatGPT. Write your own paper or create your own infographic instead of asking AI to do it.
Opt out of AI as much as possible. Don’t use Google and Meta AI, for example.
Stay educated and aware of these things and spread the word. The more people that limit our dependence on AI, the better.
If you fall for ads, find a way to turn them off so you don’t fall victim to overconsumption.

And here are some things AI companies and governments should do:

Countries should standardize AI impact reports and then companies should legally have to share their environmental impacts so that we know what harm they are causing.
Tech companies should focus NOW on water and tech recycling instead of down the road when it’s more dire. They should also implement the use of recycled materials in the building of these data centers to prevent the use of harmful extraction of new resources.
Governments should incentivize these companies to use green energy.
If AI is to continue (which there is no doubt that it will), energy and water efficiency should be prioritized. Tech companies should make searches require less water and energy.

As I said in the beginning, there are good uses for AI for the environment and for climate change. I use AI myself. I use Siri to help me set reminders and timers and I use a lot of different tools in Canva to make content. Just like with other wasteful things we do like use energy and water and plastic, we must use it sparingly. Reduce our usage as much as possible to lessen our impact and the impact of AI overall.

Of course, we are not THE problem for using AI a handful of times, but it still matters if we avoid it and use something AI-free in its place.

AI has its place, but we must be mindful and careful about how we use it. I’m sure I sound like someone from the 1920s who was scared about cars. But, look, they had a gigantic environmental impact. It’s okay to partake in this new technology while also being weary about it, especially as environmentalists.

Thank you so much for reading along. I hope that this was valuable and educational. If you have more information, please let us know below.

As always, remember that your small actions make a big difference in the long run :)

Emma

Is AI *Really* Killing the Planet?