Classroom resources Responsible
How AI uses our drinking water (BBC)
This video explores how rapid AI growth is placing pressure on global water supplies through data centre cooling and indirect water use linked to energy and hardware manufacturing.
- Lesson time
- 16 min
- Fits
- Tutor Time (15 min)
- Format
- Video
- Level
- Beginner
- Views
- 402
402 views
Before you teach
Get ready
- Test YouTube playback works on your classroom network before the session.
- Optional: have one image or article about data centres ready for extension.
Teacher notes
Sensitivity: Some students may know of local protests or water stress in their area. Keep discussion evidence-based and respectful.
Nuance: The video stresses risks. Also note the industry efforts toward efficiency and innovation mentioned near the end — without greenwashing.
Follow-up questions to push thinking further:
- Does knowing about the water cost change how you use AI? Why might individual behaviour change matter less than company accountability?
- Thames Water covers London and the Thames Valley — the same area as major UK data centre hubs (Slough, Reading, West London). Thames Water is already in financial difficulty with ageing infrastructure. How does that local angle change the debate?
- The real water cost is in training large models, not individual queries. If that is true, who should be making decisions — users, companies, or governments?
Learning objectives
- Identify the direct and indirect ways AI technology consumes water resources (approx. 1:36-4:38).
- Explain liquid cooling in data centres and why it often requires high-quality water (approx. 2:34-3:16).
- Evaluate AI infrastructure environmental impact in the context of global water security and sustainability (approx. 3:30-5:07).
The lesson
-
Starts at 0:00
Play from the start. Pause at approximately 0:40.
Teacher tip
Ask before playing: What are some everyday examples of using an AI chatbot? — activates prior knowledge.
-
Starts at 2:00
Play 2:01-3:16 (servers, why chips get hot, liquid cooling). Pause.
-
Starts at 4:00
Cold-call pairs: Why do data centre chips need cooling — and what does that have to do with water?
Teacher tip
Target answer: chips generate heat, liquid cooling removes it using water, much of that water evaporates (up to 80%) and leaves the local water circuit.
-
Starts at 7:00
Play 3:43-4:38 (electricity, direct vs indirect water use). Pause.
-
Starts at 9:00
Pairs: Give one direct and one indirect way AI uses water. Share back.
Teacher tip
Direct = cooling water at the data centre. Indirect = power plants use water to generate electricity for AI.
-
Starts at 12:00
Play 4:39 to end (transparency, company actions, innovations).
-
Starts at 13:00
Plenary: Why is it hard to get clear data from tech companies on AI water use — and what should they be required to measure and publish?
Teacher tip
Push toward accountability language: transparency, independent verification, reporting standards. This is genuinely unresolved policy territory.
Big question
If AI interactions carry a water and energy cost, how should that change when we choose to use AI — and what we ask it to do?
Adapting the lesson
Support
For pupils who need more help
Common misconceptions to address:
- The cloud has no physical impact: the cloud runs on physical data centres using electricity and water.
- Only sending a prompt uses water: impacts occur across the whole lifecycle including manufacturing and power generation.
- Cooling water is recycled forever: a large share can evaporate depending on site design.
Provide a key vocab card and sentence stems: AI uses water because… / One risk to communities is…
Stretch
For pupils ready to go further
Compare two claims in the video (e.g. Sam Altman's 1/15th teaspoon figure vs the study data). What is uncertain? Write a one-paragraph critique of how we evaluate statistics from companies with a financial interest.
SEND
For additional needs
Key vocabulary card with simple definitions and images of a data centre. Option to respond with a labelled diagram (data centre to cooling to water source) instead of written paragraphs.
In every lesson
Debate it
For the National AI Conversation
Turn it into a motion: reword the big question as "This house believes…". Split the class for and against, give sides five minutes to prepare, hear two speakers each, then take a free vote.
National AI Conversation
Set up the debate
The same topic, pitched for each age group. Pick your pupils' pathway: the motion, a prompt to get them going, and three points each way to start their preparation.
Primary · Years 5 and 6
The motion
Every time you ask AI something, it uses electricity and water. Should the app tell you how much?
Sentence starter: I think it should / shouldn't because ___.
For
- If we knew the cost, we could choose not to ask AI silly questions.
- Food labels tell us what is inside, so AI apps should tell us what they use.
- It could help save water for the people and nature living near data centres.
Against
- The numbers might be hard to understand and just worry people.
- One question uses only a tiny bit, so knowing might not change anything.
- It is the companies' job to fix the problem, not children's.
- In class
- Pupils move to a corner of the room or vote with a show of hands, then tell a talk partner why.
- As a full debate
- Shorter speeches, with more time for questions from the floor and the final vote.
Secondary · Years 7 to 11
The motion
Should people use AI less because of its water and energy cost, or is it up to the companies to fix the problem?
Challenge card: What if using AI for a task saves more energy than doing it another way?
For
- Millions of unnecessary prompts add up to real water used by data centres.
- Waiting for companies to act could take years while local water stress gets worse.
- Thinking before we prompt builds better habits and better questions.
Against
- Most of the cost comes from training models and running data centres, which users cannot control.
- Blaming individuals lets companies avoid their responsibility.
- AI can also save resources, for example by managing energy grids and finding leaks.
- In class
- Small groups take talking roles, then test their view with a “What if…?” challenge card.
- As a full debate
- Three-minute opening cases, two-minute rebuttals, and questions from the floor.
Post-16 · Sixth form and college
The motion
This house would require AI companies to show the environmental cost of every request.
The tension: Transparency vs practicality. Research: How reliable are current estimates of AI's energy and water use?
For
- Without published figures, users, regulators and local communities cannot hold companies to account.
- Labels change behaviour: energy ratings pushed appliance makers towards efficiency.
- Communities near data centres, such as the Thames Valley, deserve to know how much local water is used.
Against
- A figure for every request is hard to calculate honestly and gives false precision.
- Annual reporting for each data centre or company would be more useful and more reliable.
- A focus on single requests distracts from the larger cost of training models and building hardware.
- In class
- A formal debate, with points of information offered during the opening speeches.
- As a full debate
- Longer opening cases, and a show of hands or a digital vote at the end.
Key words
- Data centre
A large facility housing many servers. Keeping them cool requires enormous energy and often water-based liquid cooling.
- Liquid cooling
A method of removing heat from servers using circulating water or coolant. Much of this water can evaporate and leave the local water supply.
- Direct water use
Water consumed on-site at a data centre, primarily for server cooling.
- Indirect water use
Water used upstream — e.g. power stations need water to generate electricity for AI. Energy demand creates water stress before a drop touches a server.
- Water stress
A condition where demand outstrips available supply in a region — worsened when large facilities compete with communities for the same local sources.
Take it further
- Homework Find one personal habit that reduces unnecessary AI use and write one question you would ask a tech company about how they measure and report their water use.
- Next lesson Research one data centre location where local communities have protested about water or energy use. What were their concerns and what changed?
- Next lesson Thames Water serves London and the Thames Valley — the same region as some of Europe’s densest data centre corridors (Slough, Reading, West London). Thames Water is already under severe financial and infrastructure stress. Should data centres in that region face stricter water-use limits? Find one argument for and one against.
- Independent Does knowing each AI interaction has a small water and energy cost change how you think about using AI? Write 3-5 sentences explaining whether it changes your behaviour — and why or why not. (Hint: think about who is really responsible — the user or the company building the system?)