The Teen Zine 

Artificial intelligence has quickly become part of everyday life. Whether people realize it or not, AI is already involved in many of the technologies we use, from search engines and business tools to education, medicine, and scientific research. But as AI continues to grow, so does a question that cannot be ignored:

How much is AI costing the environment?

The environmental impact of AI is complicated. AI requires data centers, powerful computer hardware, electricity, cooling systems, water, land, and other resources to operate. At the same time, AI has the potential to help solve environmental problems, improve medicine, support education, and accelerate scientific research.

So, is AI worth its environmental cost?

Right now, I believe the answer is yes.

But that does not mean AI should be allowed to grow without limits. If we want AI to remain a useful technology for the next 50 years, we need to understand its environmental impact now and work toward fixing the problems that come with it.

The Environmental Cost We Can’t Ignore

When I think about AI’s environmental impact, the first thing that comes to mind is water.

There is often a lot of discussion about how much water data centers use to cool the equipment needed to operate AI systems. Water is everywhere, and about 71% of Earth’s surface is covered by it, but that does not mean all of that water is available for people or industry to use.

What matters is where the water comes from, how much is being used, and what happens when the resources available to a community become limited.

The United Nations University Institute for Water, Environment and Health recently examined the water, land, and carbon footprints associated with AI’s electricity use. Its 2026 report emphasizes that AI is not simply a digital technology floating somewhere in the internet. It depends on physical infrastructure, including data centers, advanced chips, cooling systems, electricity grids, water resources, land, and critical minerals.

The report projects that by 2030, the global data centers powering AI could consume around 945 terawatt-hours of electricity. It also estimates that the associated water footprint could equal the basic annual domestic water needs of 1.3 billion people in Sub-Saharan Africa.

Those numbers are significant, but they also need context.

AI is not the only industry using water. Agriculture, manufacturing, electricity production, households, and countless other industries depend on water. Because of that, we cannot put every water shortage on AI. A temporary water shortage, for example, does not automatically mean AI caused it.

The question becomes much more serious if environmental pressure becomes permanent.

If communities reached a point where clean drinking water was consistently unavailable, tap water could no longer be safely relied upon, or people had access to only extremely limited amounts of water, I would have to say that something had gone too far.

At that point, we would need to look not only at AI, but at every industry contributing to the problem.

That is the difference between recognizing a problem and finding a scapegoat.

Electricity Is Another Piece of the Puzzle

Water is my biggest concern, but electricity comes second.

The International Energy Agency estimates that data centers consumed around 415 terawatt-hours of electricity in 2024, approximately 1.5% of global electricity consumption. In its base-case projection, global data-center electricity consumption could more than double to around 945 terawatt-hours by 2030.

AI is one of the major reasons for that growth. The IEA projects that electricity consumption from accelerated servers, which are primarily driven by AI adoption, will grow by about 30% annually through 2030 in its base case.

That sounds alarming, and it should make us pay attention.

But it is also important not to take one statistic and turn it into the entire story.

Even with this growth, the IEA projects that data centers will account for just under 3% of global electricity consumption in 2030. In other words, AI and data centers are becoming a significant source of electricity demand, but they are not responsible for all of the world’s energy use.

This is why I think the environmental conversation surrounding AI needs more nuance.

We should be asking where the electricity comes from, how efficiently it is produced, and what infrastructure is being built to provide it. Communities can also have legitimate concerns when major energy infrastructure or data centers are proposed near where they live.

Those concerns deserve to be investigated rather than dismissed.

Efficiency Is Helpful, But It Isn’t the Entire Solution

One argument for dealing with AI’s environmental footprint is simply to make AI more efficient.

That makes sense, but I don’t think efficiency completely solves the problem.

If AI becomes cheaper, faster, and more efficient, people may simply use it more.

A solution to one problem could create another problem within the same system.

This is something we need to understand before assuming that technological efficiency automatically means lower environmental impact overall.

The OECD has emphasized that AI’s environmental footprint should be measured across its entire lifecycle, including the production, transportation, operation, and end-of-life stages of the hardware that supports AI. These impacts can include energy use, water consumption, greenhouse gas emissions, and raw-material use.

That means we should not only ask:

How much electricity does this AI model use?

We should also ask:

Where did the hardware come from?

What resources were required to produce it?

How much water does the system require?

What happens to the equipment when it is no longer useful?

And what happens when millions or billions of people begin using the technology?

Those questions give us a much better picture of AI’s environmental footprint.

But What About What AI Can Give Us?

This is where the conversation becomes difficult.

It is easy to list AI’s environmental costs. It is harder to measure what society could gain from the technology.

AI has given people resources that were previously more difficult to access.

I think of it as Google on steroids.

Search engines can help people find information, but AI can also explain information, organize ideas, answer questions, and assist people with different tasks.

AI is also becoming increasingly relevant to businesses. It can help people develop ideas, analyze information, and complete certain tasks that might otherwise require more time or resources.

Education is another example.

Some schools have banned AI completely, while others are beginning to teach students how to use it responsibly. In my own experience, I have seen teachers allow AI for certain assignments as long as students disclose that they used it.

I prefer approaches like this because instead of simply banning a technology, we can teach people how to work with it.

There is a difference between using AI to replace your own thinking and using it as a tool while remaining responsible for your work.

Medicine and scientific research could also become major areas where AI provides benefits. As someone interested in biomedical engineering, I think about what could happen if AI helps researchers identify patterns in medical data, analyze medical images, or detect diseases earlier.

Some AI applications are already being explored in scientific and environmental fields, and the OECD identifies applications such as climate prediction, environmental modeling, smart grids, and digital twins as areas where AI can contribute to sustainability goals.

That potential matters.

If AI can help us understand diseases, improve energy systems, study ecosystems, or solve scientific problems faster, then simply looking at its environmental cost without considering its benefits gives us only half of the picture.

AI Should Not Get a Free Pass

However, believing that AI’s benefits currently outweigh its environmental costs does not mean I think AI companies should be allowed to do whatever they want.

AI needs stipulations.

Powerful technologies need rules, especially when they can create serious consequences for people or the environment. The same principle should apply to AI’s environmental footprint.

The OECD has argued that policymakers need better ways to measure AI’s environmental effects and should look beyond energy and carbon emissions to factors such as freshwater use and critical mineral extraction. It also recommends greater transparency and attention to equity.

I think responsibility should be shared.

AI companies know how their systems work and how much infrastructure they require. Governments have the ability to create rules that make sure companies operate responsibly. Energy companies provide much of the infrastructure AI depends on, so communication between the technology and energy industries matters.

Consumers also play a role because the more people use these systems, the more demand they create.

Nobody should carry all of the blame.

If we don’t want water usage or electricity demand to increase, simply blaming AI companies does not solve the problem.

If we want to keep using AI, then everyone involved should contribute to making it more sustainable.

Don’t just talk about the problem. Work on it.

Renewable Energy Could Be Part of the Answer

One solution I believe deserves serious attention is renewable energy.

If companies are going to continue expanding AI, they should work toward powering that growth with renewable energy. This would not solve every environmental problem associated with AI, but it could address part of the electricity problem while encouraging companies to invest in cleaner infrastructure.

The IEA projects that renewables will supply about half of the growth in global data-center electricity demand through 2035, while other sources, including natural gas and nuclear power, will also contribute.

Renewable energy could also become part of a larger effort to rethink how AI infrastructure operates.

We need to think about electricity, cooling, water, and location together rather than treating each problem separately.

That is important because solving one environmental problem can sometimes create another.

The United Nations University report points out that low-carbon electricity is not automatically low-water or low-land. Different energy sources can have different environmental footprints.

There is no magic switch that makes AI completely environmentally harmless.

There is, however, room to make it better.

We Need to Understand AI Before We Decide Its Future

I don’t think we have enough evidence yet to say that AI’s environmental benefits definitively outweigh all of its environmental costs forever.

Right now, I believe the benefits outweigh the costs.

But that could change.

If environmental damage became severe and permanent, especially if AI became a major contributor to irreversible shortages of essential resources, I would support much stronger limits.

And I would not stop at AI.

We would need to examine every industry contributing to those shortages.

Until then, I think the better answer is not to stop AI.

It is to understand it.

The OECD describes AI as having both direct environmental impacts from the infrastructure that powers it and indirect impacts from what AI applications actually accomplish. Those indirect effects can be positive or negative. AI can contribute to environmental problems, but it can also support solutions such as smart-grid technology and environmental modeling.

That is why the question should not simply be:

“Is AI good or bad for the environment?”

The better question is:

How do we make AI good for the future?

If we are going to keep using AI for the next 50 years, we need to understand it and persevere with it.

If we can understand how something works, we can understand how to fix it, how to adjust it, what it means, and what it doesn’t mean.

We need to comprehend technology not only so we can react to it, but so we can be knowledgeable about it.

AI is not going away simply because we are uncomfortable with its environmental costs.

And ignoring those costs will not make them disappear.

The answer is somewhere between those two extremes.

We should measure the damage.

We should acknowledge the benefits.

We should create reasonable rules.

We should invest in renewable energy and more sustainable infrastructure.

We should hold companies, governments, energy providers, and consumers accountable for the roles they play.

Most importantly, we should work with the technology rather than against it.

AI has the potential to do great things, but potential is not permission to ignore consequences.

If we want AI to be a technology that lasts, then sustainability cannot be something we think about after the technology has already changed the world.

It has to be part of the plan from the beginning.

​

​

​

Sources
1.    International Energy Agency (IEA). Energy and AI: Energy Demand from AI. 2025.
This source supports the article's discussion of data-center electricity consumption, projected growth through 2030, AI-driven electricity demand, and the role of renewable energy. 
IEA: Energy Demand from AI 
2.    Organisation for Economic Co-operation and Development (OECD). Measuring the Environmental Impacts of Artificial Intelligence Compute and Applications: The AI Footprint. 2022.
This source provides the framework for examining AI's direct environmental impacts, including energy, water, hardware production, and disposal, as well as the potential positive and negative environmental effects of AI applications. 
OECD: The AI Footprint 
3.    Organisation for Economic Co-operation and Development (OECD). AI Compute.
This source supports the discussion of AI's physical infrastructure, freshwater use, critical minerals, sustainability, and potential environmental applications such as climate prediction and environmental modeling. 
OECD: AI Compute 
4.    United Nations University Institute for Water, Environment and Health (UNU-INWEH). Environmental Cost of Artificial Intelligence: Carbon, Water, and Land Footprints. 2026.
This source supports the article's discussion of AI's water, land, carbon, electricity, data-center, and infrastructure footprints, as well as the importance of environmental justice and responsible AI development. 
UNU-INWEH: Environmental Cost of Artificial Intelligence 
5.    United Nations University Institute for Water, Environment and Health (UNU-INWEH). Rising Emissions, Depleting Water and Vanishing Land: UN Scientists: AI Is Threatening Natural Resources for Billions. 2026.
This source provides the 2030 projections discussed in the article concerning data-center electricity, water, and land footprints. 
UNU-INWEH: AI's Environmental Costs