In recent years, the rapid expansion of artificial intelligence (AI) has brought major advances in computing, automation, and data analytics—but also significant environmental costs. These costs include energy expenses to maintain data centers and the carbon emissions associated with their operations as well. A typical large data center that supports AI models can consume millions of gallons of fresh water annually for cooling—comparable to the yearly water use of a town with 30,000 residents (Leppert 2025). This means that while an AI model answers your question in seconds, behind the scenes, cooling systems may be drawing on the same volume of water needed to sustain an entire community. In drought-prone regions, this trade-off poses serious sustainability concerns. This is just one example of the environmental impact. As science educators, we can help students understand both the benefits and the “hidden” ecological footprint of emerging technologies.

What Powers AI?

AI systems rely on vast networks of data centers, specialized chips, and intensive computing processes that require enormous amounts of energy, water, and raw materials. When we think of AI, we often imagine an app or digital assistant—but behind the scenes lies a physical infrastructure that consumes substantial environmental resources (Friedmann and McCormick 2024).

Key Environmental Concerns

1. Energy Consumption and Carbon Emissions
Data centers are among the fastest-growing sources of electricity demand worldwide. The International Energy Agency projects that data center electricity use—driven largely by AI—could more than double by 2030 (IEA 2024). For instance, training a large generative AI model can consume up to 1,287 MWh of electricity, equivalent to powering about 120 U.S. homes for a year, and generate more than 550 tons of CO₂ (Zewe 2025). Although the number of days/weeks/months are not provided in terms of how long the training process took, it definitely can be said that this is an aggregated energy consumption over the entire process, not a per-hour rate or duration.

2. Water Use for Cooling
AI’s environmental footprint also includes the vast amounts of fresh water used to cool servers. For every kilowatt-hour of electricity consumed, approximately 2 liters of water may be required for cooling (Zewe 2025). In water-stressed regions, this can create tension between technological and ecological priorities (Wikipedia 2025). Unfortunately, this can become a vicious cycle because the fresh water must be treated, adding to the environmental impact. 

3. Embodied Emissions and E-Waste
Beyond the operational footprint, there are upstream and downstream effects—from mining rare Earth materials for chips to disposing of electronic waste (Friedmann and McCormick 2024) from either old or inadequate chips, as well as slower boards in data centers.

4. Social and Environmental Justice
Data centers are often built in low-income or rural areas, concentrating environmental burdens such as heat, noise, and water depletion in vulnerable communities (Medium 2023). In addition, these communities are typically unaware that these centers are coming, let alone uninformed of the strain on local natural resources.

5. AI as Both Problem and Solution
AI can also help mitigate climate change by optimizing renewable-energy grids, monitoring deforestation, and modeling atmospheric systems (Bartczak and Block2025). The key is balancing technological innovation with environmental stewardship.

Why This Matters for Science Teachers

Teaching about AI’s environmental impact directly connects to NGSS performance expectations, such as HS-ESS3-4 (Evaluate or refine technological solutions that reduce the impacts of human activities on natural systems). It also develops systems thinking and digital literacy, helping students understand that “virtual” technologies have very real physical consequences.

Educators can guide students to measure, compare, and draw inferences about AI’s resource use, turning abstract global issues into authentic, data-driven science investigations. Science teachers may consider using the following lesson in their classes to demonstrate this standard. 

Example Lesson. The Carbon Cost of Our Clicks

Essential Question. What is the environmental cost of our digital habits?

This first activity turns abstract concepts like “energy use” and “carbon emissions” into measurable, relatable experiences. Students track and compare the electricity consumption of common digital activities—streaming a video, running a web search, or using an AI chatbot—and estimate how much carbon dioxide each one produces.

Using simple formulas, students multiply device wattage by time (in hours) to find kilowatt-hours (kWh), then apply a carbon intensity factor (for example, 0.45 kg of CO₂ per kWh). By comparing these activities side-by-side, they can visualize which tasks have the largest environmental footprint.

Classroom discussions often reveal surprises: Students may assume that streaming or gaming has a greater impact than AI, but when they calculate how often AI models must process enormous datasets, the scale becomes clear.

This lesson strengthens data literacy, systems thinking, and evidence-based reasoning—core NGSS science practices. It also sparks conversations about renewable energy, digital citizenship, and everyday sustainability choices.

Hypothesis/Inference Prompt. “From my data, I infer that streaming one 30-minute HD video consumes nearly twice the energy of one AI chat because the data transfer requirements for one AI chat are significantly less. 

Potential Objectives

Students will be able to estimate the electricity consumption (kWh) and CO₂ emissions of different digital tasks (streaming, web searches, and AI queries). Students should then be able to compare their results and draw inferences about the environmental costs of AI and data usage.

Materials

For this lesson, the materials needed are computers or tablets, power meter or average device wattages, a “Digital Task Tracker” worksheet, and data from your Regional CO₂-per-kWh conversion chart (e.g., 0.45 kg CO₂/kWh).

Procedure

Consider this as a procedure for this lesson.

• Introduce the issue. Explain that global data centers consumed about 460 TWh of electricity in 2022 (Zewe 2025). Ask, “What happens environmentally each time you send a prompt to an AI tool?”

• Students measure use. Track the time spent on three digital activities (e.g., 30 minutes of AI chatting vs. video streaming).

• Calculate energy. Multiply wattage × time (hours) = Wh; convert to kWh. (For example, a student uses a laptop that draws 50 watts [W] while running a 30-minute [0.5 hour] AI chat session.)

• Estimate emissions. Multiply kWh × regional CO₂ factor = estimated CO₂ output.
  
  
• Compare and infer. Students analyze which activities have higher footprints and hypothesize why.

Discussion Questions for Critical Thinking

When students compare the energy use of different digital activities, they quickly discover that tasks involving high data movement—such as streaming HD video or running multiple AI prompts—are among the most energy-intensive. Yet the environmental impact of these activities can shift dramatically when renewable electricity is part of the equation; the same number of kilowatt-hours drawn from solar or wind power generates far fewer emissions than electricity from fossil fuels. This invites students to think critically about how both personal behavior and technological design influence sustainability. Small shifts—like lowering video resolution, reducing unnecessary cloud activity, optimizing device settings, or choosing platforms committed to renewable energy—can collectively reduce the digital footprint of modern computing. Consider these discussion questions: Which tasks are most energy-intensive? How does renewable electricity change the outcome? What behavioral or design changes could lower the impact?

Data Inference Example

“From my data, I infer that streaming one 30-minute HD video consumes nearly twice as much energy as one AI chat because of sustained data transfer requirements.”

Another lesson idea could center around water use for cooling towers and the hidden footprint of AI data centers. An essential question for this lesson is How much water does it take to keep AI running?

By engaging in evidence-based measurement and comparison, students learn that AI is not immaterial: Its environmental footprint involves complex systems of energy, water, and materials. These lessons encourage young scientists to weigh innovation against sustainability, echoing the NGSS emphasis on engineering solutions for environmental challenges.

AI is transforming every field, but its environmental impact reminds us that every byte has a cost. Science classrooms are the perfect place to prepare the next generation to make informed, ethical decisions about that cost.

Connecting Science, Technology, and Ethics

Both lessons highlight that AI’s impact extends beyond algorithms: It’s a systems-level issue involving energy production, resource management, and social responsibility. By helping students calculate and interpret real data, teachers can guide them to draw informed inferences about how technology shapes the environment.

These investigations align with NGSS HS-ESS3-4 (“Evaluate or refine technological solutions that reduce impacts of human activities on natural systems”) and MS-ESS3-3 (“Apply scientific principles to design a method for monitoring and minimizing human impact on the environment”).

Ultimately, teaching the environmental dimensions of AI empowers students to ask the right questions about the technologies shaping their world.

• What are the trade-offs behind convenience?
• Does this question require an AI chat, or is a simple web search sufficient?
• How can innovation be balanced with sustainability?
• What role do we play in reducing the digital footprint of the tools we use every day?

AI might be transforming society, but science classrooms are where students learn to transform awareness into action.

References

Bartczak, J., and S. Block. 2025, June. How AI use impacts the environment. World Economic Forum. www.weforum.org/stories/2025/06/how-ai-use-impacts-the-environment.

Friedmann, J., and C. McCormick. 2024. Understanding the carbon footprint of AI and how to reduce it. Carbon Direct. www.carbon-direct.com/insights/understanding-the-carbon-footprint-of-ai-and-how-to-reduce-it.

International Energy Agency. 2024. Electricity 2024: Analysis and forecast to 2026. IEA. www.iea.org/reports/electricity-2024.

Leppert, R. 2025, October 24. Pew Research Center. What we know about energy use at U.S. data centers amid the AI boom. www.pewresearch.org/short-reads/2025/10/24/what-we-know-about-energy-use-at-us-data-centers-amid-the-ai-boom.

Medium. 2023, May 14. Data centers and the environmental footprint of artificial intelligence: What cities can do now. https://medium.com/urban-ai/data-centers-and-the-environmental-footprint-of-artificial-intelligence-what-cities-can-do-now-0b043a8feb28.

Stax Engineering. 2023. The environmental impact of data centers. www.staxengineering.com/stax-hub/the-environmental-impact-of-data-centers.

Wikipedia. 2025. Water consumption of AI data centers.  https://en.wikipedia.org/wiki/The_water_consumption_of_AI_data_centers.

Zewe, A. 2025, January 17. Explained: Generative AI’s environmental impact. MIT News. https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117.

Valerie Bennett, Ph.D., Ed.D., is an Assistant Professor in STEM Education at Clark Atlanta University, where she also serves as the Program Director for Graduate Teacher Education and the Director for Educational Technology and Innovation. With more than 25 years of experience and degrees in engineering from Vanderbilt University and Georgia Tech, she focuses on STEM equity for underserved groups. Her research includes AI interventions in STEM education, and she currently co-leads the Noyce NSF grant, works with the AUC Data Science Initiative, and collaborates with Google to address CS workforce diversity and engagement in the Atlanta University Center K–12 community.

Christine Anne Royce, Ed.D., is a past president of the National Science Teaching Association and currently serves as a Professor in Teacher Education and the Co-Director for the MAT in STEM Education at Shippensburg University. Her areas of interest and research include utilizing digital technologies and tools within the classroom, global education, and the integration of children's literature into the science classroom. She is an author of more than 140 publications, including the Science and Children Teaching Through Trade Books column.

This article is part of the blog series From Chalkboards to AI, which focuses on how artificial intelligence can be used in the classroom in support of science as explained and described in A Framework for K–12 Science Education and the Next Generation Science Standards.

The mission of NSTA is to transform science education to benefit all through professional learning, partnerships, and advocacy.