In short
Start with the idea that every AI answer runs on real computers that use electricity and water, then use a few reliable figures and a hands-on activity, such as comparing text with image generation. Link it to computing, science and geography, and keep it balanced: the aim is to help pupils use AI thoughtfully, not to make them feel guilty for using it.
Steps
- Explain that AI runs in data centres that use electricity, water and hardware.
- Use two or three reliable figures, saying who published them and how sure they are.
- Run one hands-on activity, such as comparing the cost of text, image and video requests.
- Link the topic to existing computing, science or geography lessons.
- Finish with pupils agreeing a few sensible habits for their own AI use.
Make it concrete, use a few trustworthy figures, and keep it balanced. Pupils use AI already; understanding that it runs on real machines helps them use it well.
What should pupils understand?
- AI runs somewhere. Every answer is computed in a data centre that uses electricity, often water for cooling, and hardware that has to be made and eventually recycled.
- Different requests cost different amounts. Google reports 0.24 Wh of energy and 0.26 mL of water for a median text prompt to its Gemini apps. Images cost far more: in tests by Luccioni, Jernite and Strubell, image generation averaged about 2.9 kWh per 1,000 images against 0.002 kWh per 1,000 text classifications.
- Numbers are uncertain. Most figures come from companies themselves and are measured in different ways, which is a good lesson in itself.
- Scale matters. Small amounts multiplied by very many users add up.
Which activities work?
Our lesson ideas include comparing text, image and video requests, explaining tokens with paper, deciding where a data centre should go, and "prompt golf", where pupils get a good answer in as few words as possible.
Where does it fit?
In England it links to the computing, science and geography programmes of study; elsewhere, similar strands exist in most curricula. UNESCO's AI competency framework for students includes environmentally sustainable AI.
Common mistakes
- Guilt. Pupils switch off. Focus on good choices.
- Out-of-date figures. Epoch AI's recent estimate of about 0.3 Wh per ChatGPT query is about ten times lower than an older, widely quoted 3 Wh.
- Only the negatives. Mention where AI helps the environment too.
Start with for teachers, which has facts you can use with confidence.
Sources
- Measuring the environmental impact of delivering AI at Google scale, Google (arXiv:2508.15734) (2025)
- Power hungry processing: watts driving the cost of AI deployment?, Luccioni, Jernite and Strubell (arXiv:2311.16863) (2024)
- How much energy does ChatGPT use?, Epoch AI (2025)
- AI competency framework for students, UNESCO (2024)
- National curriculum in England: computing programmes of study, Department for Education (2013)
Related
Green AI: a guide for teachers
Plain-English background on AI's energy, water and carbon cost, classroom-safe facts with sources, policy points and common misconceptions.
Green AI lesson and activity ideas
Nine low-cost green AI activities for ages 9 to 18: energy estimates, paper tokens, data centre siting, debates and a rebound game. No paid tools.
Green AI for pupils
A short, friendly guide for young people: what green AI is, why AI uses energy and water, smart habits, and how to find out more.
How do I set school rules for sustainable AI use?
How to add environmental sense to a school's AI rules: when AI is worth using, light defaults, images and video, staff habits and teaching the why.