AI, Energy Use, and the Coffee Dilemma

Sarah Downey
May 29, 2025
4 min read

Updated: Dec 19, 2025

Why energy stats don’t tell the whole story - and what mindful AI use could look like

Earlier this year, I wrote about why I still say please and thank you to AI.

Not because the machine cares. But because I do.

These small rituals remind me who I am, and how I want to be. Especially as our world becomes more saturated with machine interactions, I don’t want to lose my humanity in pursuit of efficiency. I want to keep showing up with intention.

But lately, a different hot topic has entered the conversation: AI and energy use.

Some folks have started saying they avoid AI altogether because it’s too energy intensive. And at first glance, the headlines back them up. One estimate puts the energy cost of a single ChatGPT query at about ten times that of a Google search. That sounds like a lot.

Until you zoom out.

Brew a coffee, or run 80 AI queries?

Let’s bring some perspective.

Below is a table comparing the average carbon emissions of common activities to ChatGPT queries. The CO₂ emissions data presented is based on averages compiled from multiple independent studies and reports.

Activity	Average CO₂ Emissions (grams)	Equivalent ChatGPT Queries
ChatGPT Query	3.24	~1
Watching Netflix (1 hour)	170.33	~53
Brewing a cup of coffee	263.42	~81
Washing machine (single use)	275	~85
Microwave (5 minutes)	400	~123
Electric oven (typical use)	558.33	~172
Hot shower (10 minutes)	1533.33	~473
Return flight: Toronto ↔ Vancouver	627550	~193688

These figures were cross-referenced with data from organizations such as the International Energy Agency (IEA) and peer-reviewed studies to ensure a balanced representation of current estimates.

So yes, using AI consumes energy. But so do the rest of our choices. We should be comparing it to the broader backdrop of our total energy consumption. From coffee to flights to screen time, AI is just one slice of the picture. If we're focused on caring about personal energy consumption, we need to be looking at all the ways we use energy and ask: are we spending it on what matters most?

Here’s a cheeky example: the energy used to brew a cup of coffee is roughly equivalent to running 81 ChatGPT queries. If you’re grabbing that second cup to power through a late-night project, but ChatGPT could’ve helped you finish earlier with a clearer outline… maybe the coffee isn’t the hero. And maybe AI deserves a fairer spot in our energy awareness conversation.

I know. We love our coffee.

But you see the point.

Use ≠ Waste

What’s worth reflecting on is how we use the tools available to us.

Are we just using AI for our entertainment? Are we asking vague prompts and accepting long, repetitive answers? Are we treating it like a novelty? Or, are we using it to create something more meaningful?

In my article on AI and human creativity, I talked about how these tools, used with intention, can free up space, not just for productivity, but for presence. For imagination. For that quiet, human aliveness that gets squeezed out by nonstop busywork.

If AI helps you reclaim 30 minutes of deep focus…

If it nudges an idea you might have dismissed…

If it lets you log off earlier and go walk by the water or sketch out a mind map with pen and paper…

Then maybe that’s not waste at all.

Maybe that’s the beginning of a different rhythm. One that values clarity, curiosity and being awake to the world again.

Try this prompt for more precise AI response

This morning, I asked ChatGPT to read a link to an article on book recommendations and summarize it. It couldn’t. But instead of saying “I can’t access links,” it gave me 10 unrelated book recommendations. I asked again. It gave me 10 more.

What I wanted: a simple “Sorry, I can’t do that.”

What I got: paragraph after paragraph I didn’t ask for.

If we’re concerned about energy, maybe one place to start is by teaching the AI how to respond.

You can add this under “How would you like ChatGPT to respond?” in your custom instructions:

"If a request cannot be fulfilled (e.g., a link cannot be accessed), state that clearly and briefly without offering additional guesses or suggestions."

It’s a small tweak, but over time, it reduces unnecessary processing (and your own cognitive load). A win-win.

From Panic to Principles - and the Bigger Picture

If you're cutting back on AI because of energy use, I get it. We’re all trying to live more aligned with our values. But don’t let media headlines guilt you into reactive decisions. Take a breath. Check the numbers.

Ask yourself:

Does this tool genuinely serve a purpose?
Am I clear in what I’m asking?
Is this use intentional or idle?

If we want sustainable technology, we need better systems and not just shorter prompts. And yet, how we speak to machines still matters. Because it trains how we speak as humans. That’s why I’m still saying please and thank you.

And while this article focuses on personal use, it’s also worth remembering the broader context: Most of AI’s energy impact comes from corporate-scale infrastructure and large model training. Individual choices matter, but it shouldn’t eclipse the responsibility of major tech players. That’s a bigger conversation, and one worth exploring more deeply.

References

About

Sarah Downey Sarah Downey is a Canada-based consultant helping nonprofits adopt AI safely, ethically, and confidently through governance clarity and policy development.