AI rules and societyMiddleLesson 1955 min read

AI and the environment

AI runs in data centres that use electricity and often water. It can also help protect the environment. Here are the trade-offs and what is being done.

✦ Ask the tutor🧪 Lab · 25 min📝 Lesson test
Lesson in motion

In 60 seconds

AI and the environment

AI runs in data centres that use electricity and often water. It can also help protect the environment. Here are the trade-offs and what is being done.

1/6
In simple words
Every time you ask an AI a question, a big computer somewhere does work and gets warm, like a phone playing a game for a long time. Lots of these computers together need a lot of electricity and cooling.
AI does not float in the air. It runs in data centres: large buildings full of computers. These buildings need electricity to run and cool the machines, and many use water in their cooling systems. Making the chips and buildings also uses energy and materials.

Where the environmental impact comes from

SourceWhat happens
Training large modelsMany chips run for a long time, using a lot of electricity
Everyday useEach question uses a small amount, but billions of questions add up
CoolingMachines produce heat; cooling can use electricity and water
HardwareMaking chips and servers uses energy, water and minerals; old equipment becomes e-waste
Electricity sourcePower from coal or gas causes more emissions than power from solar, wind or hydro
Note
Estimates of AI's energy and water use vary widely and change as technology changes. Be cautious with single dramatic numbers, and check who produced them and how.

Why it matters locally

Data centres are often built near cities. In places with water shortages or stretched power supplies, a large new facility can raise questions for local communities about water, electricity and land. Many regions, including parts of India, face water stress, so planning matters.

AI can also help the environment

  • Better weather and monsoon forecasting to help farmers plan.
  • Balancing electricity grids that use more solar and wind power.
  • Spotting leaks in water pipes and waste in factories.
  • Monitoring forests, air pollution and wildlife from satellite images.
One-sided views
  • “AI is destroying the planet, so never use it”
  • “AI will solve climate change by itself”
Balanced view
  • AI has real environmental costs that should be measured and reduced
  • AI can help some environmental work, but it is one tool among many

What is being done

  1. 1

    More efficient models

    Researchers make smaller, faster models that need less computing for many tasks.
  2. 2

    Cleaner power

    Some companies buy or build renewable electricity for their data centres.
  3. 3

    Better cooling

    New cooling designs aim to use less water and energy, and some reuse waste heat.
  4. 4

    Measuring and reporting

    More pressure from researchers, customers and some regulators to report energy and water use.
  5. 5

    Smarter siting

    Placing data centres where clean power and water are more available.
Real example
Everyday example. A student in Chennai notices her city faces water shortages in summer and reads about new data centres being planned nearby. She writes a school project asking what cooling systems they will use and where their electricity will come from. Asking these questions is exactly what good public discussion needs.
Do this
You can reduce your own share: use AI when it genuinely helps, avoid generating dozens of images or long answers you do not need, and choose simpler tools for simple tasks.
Try this. Find a sustainability or environment report from one large technology company. Look for what it says about data centre energy, renewable power and water. Write down what is clearly measured and what is vague.

Watch and read more

Lab

A short, evidence-based brief on the environmental trade-offs of AI in your region.

~25 min

The problem

Write a one-page brief for a local council or school.

  1. Explain in simple words where AI's environmental impact comes from.
  2. Find one reliable source (a company report, research body or trusted news) about data centre energy or water. Note what it measures and its date.
  3. List two local concerns (for example water or power supply) and two possible benefits of AI for your region's environment.
  4. Suggest three questions the community should ask before a new data centre is approved nearby.
  5. Do not invent numbers; if you use one, cite its source.

You are done when

Hard questions

Try to answer before you reveal. If you can answer these, you understood the lesson.

Q1If AI becomes more efficient, why might its total energy use still grow?Reveal

When something becomes cheaper and easier, people often use much more of it. If each AI answer uses less energy but the number of answers grows much faster, total use can still rise. This is sometimes called the rebound effect, or the Jevons paradox. Efficiency helps, but total impact also depends on how much AI is used and how clean the power is.

Q2Why can reducing water use in a data centre sometimes increase energy use?Reveal

Cooling methods involve trade-offs. Evaporative cooling uses water but can save electricity; air-based or closed-loop systems may use less water but more electricity. The best choice depends on local conditions: in a water-stressed area, saving water may matter more; where electricity is from fossil fuels, saving energy may matter more. Good decisions look at both together.

Please sign in to continue.

Questions people ask

Does one AI question use a lot of energy?

A single question uses a small amount, and the exact amount depends on the model, the task and the data centre. The concern is mainly the total across huge numbers of users and the energy used for training.

Why do data centres use water?

Computers get hot, and many cooling systems use water, for example by evaporating it to carry heat away. Some designs use much less water, but they may need more electricity instead. It is a trade-off.

Should I stop using AI to protect the environment?

That is a personal choice. A balanced approach is to use AI when it truly helps, avoid waste, and support transparency and clean energy. Everyday choices like transport, electricity and food usually also matter a great deal.

Is anyone checking AI's environmental impact?

Researchers, journalists, environmental groups and some governments are studying it, and some companies publish sustainability reports. Methods are still improving, so figures are often debated.

Lesson test

5 questions. Get 3 right (60%) to pass and complete this lesson.

Sign in to take the test and save your progress