AI is everywhere at the moment. We are talking about it in relation to workload, planning, assessment, creativity, feedback and the future of teaching. But one question is coming up more and more often and it is one we cannot ignore, 'What is the environmental impact of AI?'
That was the focus of the latest STEM Community Live: AI Sprints session, where Alex More was joined by Rita Bateson, Stephen Taylor and Mark Langley for a thoughtful discussion about what sits behind the apparently effortless experience of using AI tools. While AI can feel instant, easy and invisible, the systems powering it are anything but.
Looking beyond the prompt
When a teacher types a prompt into a chatbot, or a student generates an image, summary or piece of text, it can feel like a simple digital interaction. In reality, that request is supported by a huge physical infrastructure of servers, data centres, cooling systems, energy supply, raw materials and land.
As educators, we should absolutely be helping young people understand how to use AI well. But we should also be helping them understand the wider systems behind it, the trade-offs involved and the questions it raises. This is not about scaremongering. It is about digital literacy in the fullest sense.
Four critical topics
The conversation was framed around four topics: energy, water, materials and land. Together, these provide a helpful way of thinking about AI’s environmental footprint.
It's easy to think of AI as something that lives entirely in the cloud, detached from the real world. But these four topics remind us that AI is rooted in physical systems and real-world resources. It uses electricity, relies on water for cooling, depends on mined materials and manufactured hardware, and takes up space physically and environmentally.
Red AI and green AI
One of the key ideas explored in the session was the distinction between “red AI” and “green AI”, drawn from the paper Green AI by Schwartz, Dodge, Smith and Etzioni (2019). In simple terms, red AI is about scale, speed and power. It is the push towards bigger models, more compute and ever greater performance, often with less attention paid to environmental cost.
Green AI asks a different question: can we develop and use AI in ways that are more efficient, transparent and sustainable? That feels especially relevant for education. Not every task needs the most powerful model, and not every classroom use of AI needs to be flashy or prompt-heavy. One of the strongest messages from the session was that purposeful use matters.
Why schools should be part of this conversation
What I liked about this discussion was that it did not fall into easy extremes. It was not an anti-AI session. It was not a sales pitch either. Instead, it sat in the space where many teachers are currently working: interested, curious, cautious and trying to make good decisions.
Educators are in a strong position here because this topic opens up rich opportunities for discussion and enquiry. Students can explore how energy is used, why data centres need cooling, where materials come from, and how innovation connects with responsibility. There is also the pastoral dimension. The session touched on the risk of layering AI anxiety on top of existing eco anxiety for young people. We want students to be informed, not overwhelmed, and to think critically without feeling powerless.
Bringing the issue down to earth
One of the most interesting parts of the session was hearing from Rita Bateson, who shared some really eye-opening insights into the spread of data centres in and around Dublin and just how much electricity they are pulling from the Irish grid. It brought the whole issue much closer to home. Questions around energy demand, cooling, water use and infrastructure all affect real communities. That matters, because it helps teachers and students see that AI is not just something floating around online. It has a physical footprint and real-world impacts.
A useful teaching resource
Another highlight was Stephen Taylor’s If You Use Me AI, which includes an AI carbon estimator alongside a range of useful links and prompts for further exploration. What I particularly like about this resource is that it gives teachers a practical way into the conversation. It helps make a complex issue more visible and more discussable, whether in the classroom, a STEM club or wider school conversation. (Link below)
Hope, and informed choice
AI does have the potential to support learning, widen access and help some young people engage in ways that might otherwise be difficult. But that does not remove the need for scrutiny. It increases the need for it. The challenge for education is not to reject AI outright, nor to adopt it uncritically. It is to help students and staff make informed choices, ask better questions and understand systems as well as tools.
Explore the ideas further
Here are the key links shared in and around the session:
Red/Green AI: https://arxiv.org/abs/1907.10597
The ecology of artificial intelligence: energy, water, materials, and land limits of digital systems: https://link.springer.com/article/10.1007/s44438-025-00018-8
If you use me estimator: https://ifyouuseme.ai/environmental-impacts
Thirsty Work: The truth about AI and its impact on the planet: https://open.substack.com/pub/alexmore/p/thirsty-work