A lot of people feel tension between exploring AI's creative possibilities and worrying about its environmental cost. One of our newsletter readers wrote to us:

Her concerns are valid. The media has been full of stories about AI using up scarce water resources and electricity as data centers spread worldwide. Water is used to cool the massive servers that run the AI models and of course all these new servers use electricity. Lots of electricity.

We sent the comment to Ken Kahn, author of The Learner's Apprentice: AI and the Amplification of Human Creativity, published by our own Constructing Modern Knowledge Press.

In this post, Ken tries to put that worry in perspective with real numbers.

Here are Ken's main points:

The bottom line on everyday AI use: A typical back-and-forth conversation with a chatbot like ChatGPT or Claude uses about as much energy as watching TV for the same amount of time — and roughly five drops of water. That's far less alarming than many headlines suggest.

What about writing longer things? Generating something like a college essay uses about as much electricity as running a microwave for three minutes. Not nothing, but not catastrophic either.

Location matters: Data centers draw from local water supplies — so the impact isn't the same everywhere. What's manageable in one region can be a real strain in another.

What about training the AI? Training a cutting-edge AI model uses less water than a single square mile of farmland uses in a year — and companies only do it a few times a year. It accounts for roughly 10–20% of AI's total energy use.

What uses more energy? Images, audio, and especially video generation are significantly more resource-intensive than text. That's worth knowing when choosing activities.

For privacy- or eco-conscious classrooms: There are AI tools that run entirely on a laptop or phone, with no internet needed — and Chrome now has a chatbot built right in.

The teaching opportunity: Rather than avoiding AI for environmental reasons, this is actually a great opening to teach students how individual choices — in technology and elsewhere — connect to global resource use. Awareness is the goal, not avoidance.

But wait there's more!

If this information got you thinking about how AI fits into your teaching practice — not just its environmental footprint, but its genuine potential to amplify what students can create and learn — Ken Kahn's book The Learner's Apprentice: AI and the Amplification of Human Creativity is the natural next step. Ken goes far deeper into the research and the possibilities, with hundreds of real examples of AI in creative, meaningful learning contexts.

​The Learner's Apprentice: AI and the Amplification of Human Creativity, published by Constructing Modern Knowledge Press.

It's for PD AND the classroom - two for one!

We are getting orders for classroom sets of this book — which means it's finding its way into schools both as a professional development resource for teachers AND as a text students can read and engage with directly. If you'd like to bring The Learner's Apprentice into your school or department, contact us about bulk pricing and using POs.

This email used approximately three drops of water to reach you. You're welcome. 😁

Sylvia Martinez & Gary Stager

The Language of Computation: Constructing Modern Knowledge in Reggio Emila
June 15-19, 2026 | Register today!​

The future is computational, children are natural mathematicians, and computation enriches the learning possibilities for all.