Lesson Ideas for the AI Footprint Estimator

Idea 1: What's The Impact of My Project?

This is the easiest one: if students use AI tools in their class project, ask them to: 

1. Keep track of their AI use across tools. This is easier if you manage a project in a single thread using your AI tool. 

2. At the end of the project, put the totals into the estimator and generate the impacts. 

   • Determine the Impact: What does this mean in terms of energy and carbon? 

   • Evaluate the Impact: Which uses of AI are most justified? Least justified? What could have been replaced with another, more efficient approach? 

   • Suggest an improvement plan: Based on the data, how can students have a reduced impact through more sensible approaches to their AI use? Collate the ideas and create a "planet friendly approaches to AI" suggestion wall. 

3. Extension: Visualise the class data:

   • Collate the entire class data and create different forms of data visualisation. 

Adaptation: Ask students to include an "environmental footprint" statement in their report. 

Idea 2: Thinking Routines with AI Eco Data

This builds on the data collection above. Try some of the following adapted PZ Thinking Routines to engage with the learning. Always follow-up with the question "What makes you say that?" to elaborate on thinking. 

"I used to think... Now I think.. Now I wonder...." Collect student thinking on what they thought about AI use before they engaged with the task. After completing the analysis, shift to "now I think..." - how has their perspective on AI changed? Follow-up with "now I wonder..." to generate new questions for research about the impacts of AI. This could be extended with Think, Feel, Care. 

Circle of Viewpoints. Considering the findings of the class in terms of their AI impacts, use Circle of Viewpoints to identify as many stakeholders as possible in the issue of AI and its impacts. What perspectives and positions around the use of AI and its impacts can they identify? Where can their research help them find out more? 

Options Explosion. Considering the different ways in which they have used AI, and the associated impacts, what are some other possibilities and opportunities? With AI? Without AI? With more thoughtful applications around AI? Unpack them more deeply to think about implications and possible extensions. 

Chalk Talk. Put "AI Impacts" at the centre of the diagram and give each student different coloured markers. In silence, let the routine run for as long as possible, with students making connections between each other's ideas, building extensions on their thinking and suggesting alternative perspectives or solutions. Once the silent part is done, use the map to build a class discussion. [AI adaptation: take a photo of the map and ask AI to transcribe and explain the connections; compare this to the outcomes of the class discussion.] 

Parts, Purposes, Complexities. Use this routine to get into more depth on conversations around AI and its implications. This could take directions of environmental impacts or the complexity of trying to work out these calculations. It might even venture into the complexities of how AI models are created and how they work. This could be further extended to Parts, Perspectives, Me. 

Compass Points. Use this one after calculating the impact data to break down perspectives and ideas into what they're Excited About wit AI, what they Need to Know about AI (and impacts), what they Worry About with AI, and their Stance or Suggestions moving forwards. 

Tug of War. Use this routine to place "AI Use" and "No AI" at opposite ends of the line. students place perspectives and thoughts about AI use and its impacts on learning (along the top) and environment (along the bottom). As a class discussion: How can we balance the possibilities and negative impacts of AI in a way that gives the most benefit with the least harm? 

Idea 3: Compare Your Carbon and Energy Footprints

Every choice we make has an impact: it's not just AI. Students can do deeper into research on the fooprints of their other activities. For example, the following table comes from Perplexity Deep Research:

• Students can analyse and verify the claims and sources with their own research. 

• Students could follow the sources and critically evaluate their reliability.  

1 hour scrolling on TikTok (smartphone, Wi‑Fi)

• Energy: Around 0.01–0.02 kWh per hour, combining a few watts for the phone plus a small share for networks and data centers. (Sources: “TikTok’s Global Carbon Footprint Nearly Exceeds Greece’s, Study Finds”, “Mindless TikTok-scrolling is as damaging to environment as your car, study finds”)

• Carbon: Roughly 150–180 g CO₂e per hour, based on about 2.9 g CO₂e per minute of TikTok use. (Sources: same as above plus “Is TikTok greener? Study finds carbon footprint larger than expected”)

• Water: On the order of 0.1–0.2 liters of evaporated water from electricity generation for that energy use, using global‑average power‑sector water intensity. (Sources: “Consumptive Water Use for U.S. Power Production”, “AI’s Challenging Waters”)
  
  

1 hour streaming Netflix‑style HD video (TV + internet)

• Energy: About 0.25–0.6 kWh per hour, combining roughly 0.1–0.25 kWh for streaming infrastructure with 0.17–0.34 kWh for a typical TV. (Sources: “Factcheck: What is the carbon footprint of streaming video on Netflix?”, “The Carbon Cost of Streaming”)

• Carbon: Roughly 30–100 g CO₂e per hour; central estimates are about 36 g CO₂e per hour for streaming alone, with the rest from powering the TV depending on grid mix. (Sources: same as above, plus “What Is the Carbon Footprint of Video Streaming?”)

• Water: Around 2–5 liters of evaporated water per hour from power generation for that electricity use. (Sources: “Consumptive Water Use for U.S. Power Production”, “Data Centers, Digital Lifestyles and Water Use”)
  
  

1 hour of general web search and browsing (laptop)

• Energy: Approximately 0.02–0.07 kWh per hour, dominated by a 20–50 W laptop, with only a few watt‑hours from search infrastructure (about 0.3 Wh per search). (Sources: “Our 2024 Environmental Report” (Google), “Every Google search results in CO2 emissions. This real-time data viz shows how much.”)
  
  

• Carbon: Around 15–40 g CO₂e per hour, assuming roughly 50 searches at ~0.2 g CO₂e each plus emissions from powering the laptop on a typical grid. (Sources: same as above, plus “Google searches – CO₂ Calculation”)
  
  

• Water: Roughly 0.15–0.5 liters of evaporated water per hour, based on average water use per kWh of electricity. (Sources: “Consumptive Water Use for U.S. Power Production”, “Data Centers, Digital Lifestyles and Water Use”)
  
  

1 hour online console gaming (Xbox‑class, active play)

• Energy: About 0.2–0.3 kWh per hour, with the console itself drawing around 150–220 W during gameplay plus the TV’s additional load. (Sources: “How Many Watts Does an Xbox Series X Use”, “Xbox Wattage Guide: How Many Watts Does the Xbox Series X Use?”)

• Carbon: Roughly 60–210 g CO₂e per hour depending on whether the local grid is closer to 300 or 700 g CO₂e per kWh. (Sources: same as above, plus grid‑intensity factors from national electricity statistics. 

• Water: Around 1.5–2.5 liters of evaporated water per hour from electricity generation at typical water‑use intensities. (Sources: “Consumptive Water Use for U.S. Power Production”, “Data Centers, Digital Lifestyles and Water Use”)
  
  

One beef cheeseburger

• Energy: Life‑cycle energy equivalent of roughly 2–5 kWh per burger (7–20 MJ), including feed production, processing, transport, and cooking. (Sources: “The Footprint of a Cheeseburger (Updated!)”)

• Carbon: About 3 kg CO₂e per burger, within a broader life‑cycle range of roughly 2–6 kg CO₂e depending on production system. (Sources: same as above, plus “The Climate Cost of Food at COP24”, and comparative burger LCAs. 

• Water: Around 2,000–2,300 liters of water per burger on a global‑average basis, driven mostly by feed and pasture. (Sources: “How Much Water Does it Take to Grow a Hamburger?”, “This is how much water is in your burger”)
  
  

One 12 oz black coffee (Starbucks‑style)

• Energy: Direct brewing uses about 0.02–0.05 kWh per cup; total farm‑to‑cup primary energy is several kWh equivalent when back‑calculated from emissions. (Sources: “Brewing a Sustainable Future: The Carbon Footprint of Your Coffee”)

• Carbon: Approximately 0.26 kg CO₂e per 12 oz cup. (Sources: same CDP analysis as above)

• Water: About 130–140 liters of water per cup, with most use at the farm stage. (Sources: “How big is your water footprint?”, “The water needed to have the Dutch drink coffee”)
  
  

One 12 oz latte (with dairy milk)

• Energy: Similar direct electricity use to black coffee (~0.02–0.05 kWh), but higher total primary energy because of milk production. (Sources: “Brewing a Sustainable Future: The Carbon Footprint of Your Coffee”)

• Carbon: Around 0.84 kg CO₂e per 12 oz latte, more than triple black coffee due to dairy. (Sources: same CDP analysis)

• Water: Several hundred liters of water per cup when combining the coffee’s ~130–140 liters with the high water footprint of cow’s milk (around 1,000 liters per liter). (Sources: “The water footprint of soy milk and soy burger and equivalent animal products”, “A Comprehensive Introduction to Water Footprints”, “The water needed to have the Dutch drink coffee”)
  
  

Idea 4: Investigate the Human & Ecological Impacts of AI's Environmental Costs

The huge expansion of data-centres to serve AI is having impacts on the human and ecological communities in their area. Here are a few examples; students could find out more about these, and search for other cases that connect to communities or ecosystems that are close to them. 

Chilean community art project exposes AI’s “excessive thirst”

A cultural group in one of Chile’s most water‑stressed regions created art‑driven activism to highlight how each AI chatbot interaction consumes water via data centre cooling, warning that communities like Quilicura will “suffer immensely” as AI expands and invisible water use rises.

Find out more: Quili.AI

Dutch farmers fight Microsoft over land and nitrogen

In rural Netherlands, farmers and local residents have protested Microsoft and other tech firms’ hyperscale data centres, arguing that the centres consume fertile land, add to nitrogen and carbon pressures, and receive permits and infrastructure priority while agricultural and housing projects are constrained, fueling a powerful political backlash.

Find out more: WIRED – These Angry Dutch Farmers Really Hate Microsoft

Irish communities question data centres’ grip on the grid

Local groups and national campaigns argue that the country’s dense cluster of data centres threatens ordinary people with higher blackout risk and reliance on fracked gas imports, and undermines climate targets, leading to community‑driven opposition and a de‑facto moratorium on new centers around Dublin. 

Find out more: Tech Monitor – Ireland’s data centre nightmare – and what others can learn from it

Arizona desert cities confront data centres’ water draw

In Mesa and wider Maricopa County, communities already facing groundwater limits have watched Meta, Google, and Microsoft build or plan billion‑dollar data centres permitted to use water volumes equivalent to tens of thousands of residents, sparking fears over water security for future housing and local users.

Find out more: EthicalGEO – The Cloud is Drying our Rivers: Water Usage of AI Data Centers

AI growth linked to pollution, mining, and habitat threats

Environmental advocates in the United States warn that AI‑driven data center expansion increases fossil fuel combustion, accelerates mining for critical minerals, and generates large quantities of electronic waste, all of which contribute to ecosystem degradation and threaten biodiversity across extraction sites, power corridors, and local receiving environments.

Find out more: Hoosier Environmental Council – Riddled with pollution and problems: The case for slowing AI’s growth

From forests to “cement boxes” in local backyards in North Virginia

Residents who grew up near emerging data centre clusters describe how forested land was incrementally cleared and replaced with massive windowless buildings, high fencing, harsh lighting, and constant equipment hum, transforming local landscapes and reducing green space and wildlife habitat in exchange for largely invisible digital services.

Find out more: MIT Climate & Sustainability Consortium – Investigating the Ecological Impacts of Data Centers

Idea 4: Our World in AI Data

Apply any of the routines above, based on their data, to find out more about the "bigger picture" of AI, using some of the data visualisations from Our World in Data. Start with one graph together, then use this to generate new lines of inquiry based on what is available in the datasets. Two useful starter graphs are below. 

Idea 5: Visualise This

Taking inspiration from Our World in Data, create a range of different data visualisations for the AI impacts of the class. Use this opportunity to connect to the mathematical and communication skills needed in your subject. Create a gallery walk of the visualisations, and pair with some thinking routines to learn from and give feedback to each other. 

Idea 6: Check Out Leon Furze's "Teaching AI Ethics: Environment" Resources

Leon Furze has a fantastic series on Teaching AI Ethics. This post, on Environmental Impacts, has great ideas for understanding the different levels of impact of AI and suggestions for curriculum connections. 

Idea 7: Build Your Own Data Visualisation App

If you have students who want to roll up their sleeves and get coding, set a data visualisation challenge to replicate this app or create something entirely new to calculate the impacts of their choices or other school-wide data collection. Examples might include calculating the impacts of the school cafeteria (and the effects of a "Meatless Monday"), tracking waste sorting, school-wide energy usage or more. 

Idea 8: Go Deeper Into AI Footprint Research

Using the 2026 paper "The AI Climate Hoax: Behind the Curtain of how Big Tech is Greenwashing Impacts," consider some of these critical and analytical thinking lessons. Use some of the thinking routines above. 

1. Analyse the "Bait-and-Switch" Narrative

Research the differences between traditional AI (like image classification) and generative AI (like ChatGPT), then explain why the report claims that "purported climate solutions are coupled to extreme pollution and presented as a package deal".

2. Evaluate Evidence Strength in Corporate Claims

Investigate why only 26% of the 154 climate benefit claims analysed in the report were backed by published academic research, while 36% cited no evidence at all, and discuss how this affects the credibility of big tech's "green" promises.

3. Investigate "Tactical Vagueness" as Greenwashing

Analyse the report’s concept of tactical vagueness, where companies use the umbrella term "AI" to mask impacts, and evaluate the claim that "the term AI is so vague that it’s like the term transportation... you could be talking about a bicycle or a rocket".

4. Assess the Reality of "Net-Climate Benefits"

Identify a specific corporate claim, such as Google’s assertion that AI could mitigate 5 to 10% of global emissions by 2030, and use the report’s findings to debate why "the assertion that AI’s climate benefits will outweigh harm lacks any credible basis".

5. Identify Infrastructure and Fossil Fuel Impact

Explore how the rapid expansion of data centers "rescues prospects for fossil fuels by boosting demand and triggering panicked deployment of new fossil infrastructure," and analyse how this contradicts the industry's stated climate goals.

Idea Collection: IB-Connected Engagements

This app was designed in the context of an IB World school, so here are some suggestions for how it might connect to different MYP subject areas and interdisciplinary learning. Across all ideas, there can be strong connections made with the ATL skills, particularly critical thinking, communication and transfer. 

MYP Individuals & Societies

Inquiry focus: How do technologies reshape environmental responsibility and global inequality?

Learning engagements

• Students compare AI footprints under different grid types and map these to real countries.

• Case study: “Clean grid user, coal-heavy data centre” – who bears responsibility?
  
  

• Debate: Should AI emissions be regulated like aviation or shipping?
  
  

Key concepts: Systems, Global interactions
Assessment idea: Position paper using calculator scenarios as evidence

MYP Sciences

This is ideally suited to Criterion D: Reflecting on the Impacts of Science. 

Inquiry focus: How do energy transformations and resource systems underpin digital technologies?

Learning engagements

• Use the estimator to analyse energy vs water trade-offs in AI use.

• Model uncertainty: why no “correct” footprint number exists.
  
  

• Compare AI inference energy to everyday lab equipment usage.
  
  

Key concepts: Systems, Change
Assessment idea: Scientific explanation with uncertainty analysis

MYP Mathematics

This is ideally suited to real-world applications of mathematics, communicating and investigating patterns. 

Inquiry focus: How does mathematical modelling shape our understanding of real-world problems?

Learning engagements

• Reverse-engineer calculator formulas (kWh → CO₂ → equivalencies).

• Sensitivity testing: Which variable matters most and why?

• Critique equivalencies (trees, car journeys): helpful or misleading?
  
  

Key concepts: Relationships, Representation
Assessment idea: Modelling report with justification of assumptions

MYP Design

Inquiry focus: How can design make invisible systems visible and actionable?

Learning engagements

• Redesign one calculator output for a specific audience (younger students, parents, policymakers).
  
  

• Evaluate UX decisions: task-based vs query-based inputs.
  
  

• Prototype a “low-impact AI use guide” for students.
  
  

Key concepts: Development, Function
ATL: Iterative design, Reflection
Assessment idea: Design portfolio with ethical justification

MYP Language & Literature

The application of AI in Lang&lit has more ethical considerations than just sustainability. Students could use some of the routines abocve to explore the combined implications with voice, identity and language development. 

Inquiry focus: How does language frame technological responsibility?

Learning engagements

• Analyse media claims about “AI’s carbon footprint” against calculator nuance.

• Write an op-ed using estimator data responsibly (no false certainty).

• Compare metaphor use: “AI backpack,” “hidden cost,” “digital pollution.”
  
  

Key concepts: Perspective, Communication
Assessment idea: Persuasive writing with annotated evidence use

MYP Language Acquisition

There is an opportunity here to explore how well LLM's compare across different languages or origins (for example Chinese vs English LLM in the development of Chinese fluency). 

Inquiry focus: How do multilingual technologies intersect with sustainability?

Learning engagements

• Compare AI-supported translation vs human translation footprints.

• Discuss trade-offs: access, inclusion, environmental cost.

• Vocabulary work around sustainability and technology ethics.
  
  

Key concepts: Communication, Connections
Assessment idea: Reflective comparison task

MYP Arts

With image, audio and video generation being the highest impact in terms of the environment, and of particular concern to artists, there is much to discuss here. 

Inquiry focus: What is the environmental cost of digital creativity?

Learning engagements

• Compare AI-generated images vs traditional digital workflows.

• Create artwork visualising “invisible” AI impacts.

• Artist statement must reference calculator estimates.
  
  

Key concepts: Expression, Identity
Assessment idea: Artwork + critical reflection

MYP Physical & Health Education & Social-Emotional Learning

Inquiry focus: How does digital efficiency relate to human wellbeing and sustainability?

Learning engagements

• Compare AI time-savings vs energy costs.

• Discuss cognitive offloading and environmental trade-offs.

• Link to digital wellbeing and intentional tool use.
  
  

Key concepts: Balance, Responsibility
Assessment idea: Personal AI-use audit and action plan

Interdisciplinary Unit Ideas

Unit 1: “The Invisible Infrastructure of Learning”

Subjects: Sciences + Individuals & Societies + Design
Statement of inquiry: Technological systems shape environmental outcomes through complex, often hidden, interactions.

Students:

• Map AI systems as sociotechnical systems

• Use the estimator to explore unintended consequences

• Design an intervention to reduce impact without banning AI

Unit 2: “Efficiency Isn’t Neutral”

Subjects: Mathematics + Language & Literature + Design
Statement of inquiry: Quantification and communication influence ethical decision-making.

Students:

• Model AI impacts mathematically

• Analyse how numbers persuade or mislead

• Redesign messages for integrity and accuracy

Unit 3: “Global Problems, Local Choices”

Subjects: Individuals & Societies + Sciences + Language Acquisition
Statement of inquiry: Personal decisions contribute to global sustainability challenges in uneven ways.

Students:

• Explore geographic inequity in AI infrastructure

• Debate responsibility across borders

• Propose principles for ethical AI use in schools

IBP Environmental Systems & Societies (ESS)

There are a lot of potential connections with this course. Students could use the estimator as a "tuning in" activity, and follow-up through the references to build their own understanding and research. Maybe they can make suggestions to make it better. 

Topic 2: Ecosystems & Ecology

• AI data centres as anthropogenic systems

• Water withdrawal and thermal pollution implications

• Compare digital vs physical infrastructure footprints

Topic 5: Environmental Economics

• Externalities of “free” AI services

• Cost-benefit analysis including carbon and water

• Who pays, who benefits?
  
  

Topic 7: Climate Change

• Grid intensity as a climate variable

• Demand-side mitigation through behavioural change

• Compare AI emissions to other digital services

ESS Internal Assessment (IA) Ideas

• How does AI usage pattern choice influence estimated carbon emissions under different grid scenarios?

• To what extent does AI-assisted research reduce total environmental impact compared to traditional workflows?

• How reliable are equivalency-based representations of AI environmental impacts?
  
  

Students can:

• Use the estimator as a secondary data source

• Critically evaluate assumptions and uncertainty

• Explicitly discuss limitations (which examiners love)

If you have tried any of these lessons with your students, please let me know how it went. It's best to reach me on LinkedIN.