What Your Fabric and Your AI Have in Common
Happy Earth Month, quilty friends.
Here’s an evironmental stat that surprised me when I read it.
"A single cotton t-shirt requires about 2,700 liters of water to produce. Not to wash. To make. From growing the cotton to processing and dyeing the fiber, one t-shirt uses up enough water to keep a single person hydrated for 900 days.".
Sit with that for a second. Your stash, every yard of quilting cotton, every bundle you’ve ever bought, represents water on a scale most of us have never thought about. It’s not a reason to stop quilting. It’s just a fact worth knowing (and one reason some of us keep all our scraps).
Now here’s the other number, and this one stopped me:
Training ChatGPT-3, just the training phase, before a single person ever used it, consumed roughly 700,000 liters of water, lost as steam from data center cooling towers. That’s the water needed to manufacture about 370 cars. Or, the equivalent of producing about 259 cotton t-shirts. (Source: University of California, Riverside, 2023)
And that’s just one model, trained once (ChatGPT is up to version 5.4 now). Every 20 to 50 queries you send to an AI chatbot uses approximately another half liter.
When a trusted friend challenged me on it, I spent months researching AI’s environmental impact — water, energy, noise, hardware. What I found became my research paper, Thirsty Intelligence. And the water piece is where it gets personal for anyone who works with fabric.
I think of it like this. A rotary cutter is one of the most useful tools in a quilter’s kit. It’s also the one most likely to send you to urgent care if you’re not paying attention. We don’t stop using rotary cutters. We learn to use them with intention.
AI is the same kind of tool. Powerful. Worth having. Worth respecting.
It's Not Just About Water
Water gets the headline, but the full picture in Thirsty Intelligence covers these areas:
Energy: Training large AI models burns significant electricity and produces real carbon emissions. Without changes, AI’s energy footprint could double by 2030. The good news: efficiency is improving. Google used AI on its own data center cooling and cut energy use by 15%. The technology can be part of the solution, if we push for it.
Noise: Data centers run at a constant 55–85 decibels from cooling systems alone — louder when backup generators kick in. Communities near major data center clusters have filed lawsuits over the impact on daily life. This is a physical, local problem that doesn’t show up in your query results.
Hardware and e-waste: The chips that run AI require rare earth metals — lithium, cobalt, gold — with serious upstream mining costs. Globally, 62 million tons of e-waste were produced in 2022, with only about 22% recycled. Every device upgrade feeds that number.
AI is Here. But What Does it Mean for Creativity?
If you’ve ever wondered “Is AI going to replace real art?” you’re not alone. Many creatives feel curious but cautious.
The truth? AI is a tool, not a threat. But only if you understand how it works, its potential for good and harm, and how to use it ethically and with intention.
You don’t have to dive in blindly.
You can learn the landscape, explore the possibilities, and stay firmly in control of your creative voice — while adding more speed, variety, and inspiration to your process.
Here's What Made Me Feel Better (And What Should Make You Feel Better Too)
The same AI that has an environmental cost is also being used to address environmental problems.
AI-powered irrigation systems can cut agricultural water use by around 25% while maintaining crop yields. Machine learning is being used to detect deforestation in the Amazon in near-real-time. Wildlife conservationists use AI to identify individual endangered animals from photographs without ever disturbing them.
I believe AI is “chaotic neutral” (thanks for the term, D&D). It’s not a villain, and it’s not a savior. It’s a tool that reflects the intentions of whoever is using it. Which means we get to have some say in what AI becomes.
What About Just... Not Using It?
I looked at this seriously. And the honest answer is: opting out doesn’t opt you out. AI is already embedded in spam filters, GPS navigation, medical diagnostics, online banking and more. You’re using it whether you’ve consciously decided to or not.
More importantly, avoiding AI doesn’t avoid environmental impact, it just moves it around. The research is pretty clear that a more sustainable path is intentional use, not abstinence.
Think of it the way you think about fabric sourcing. You probably haven’t stopped buying fabric. But you might have started asking where it comes from.
Five Things You Can Do Right Now
- Use AI when it earns its place. Generating colorway options, drafting product descriptions, summarizing research…those are real use cases with real returns. Idly pinging a chatbot is the digital equivalent of leaving the tap running.
- Craft better prompts. One specific, well-thought-out prompt almost always beats five vague ones, and uses fewer resources to get there.
- Choose tools that are transparent. Some AI services publish energy and water use data. Favoring those companies sends a signal.
- Hold onto your devices longer. The environmental cost of hardware is significant. An extra year on your current laptop matters more than you’d think.
- Ask for accountability. When companies know their users care about environmental impact, it moves the needle. It happened with fabric mills. It can happen here.
This Is the Part That Quilters Actually Get
We already read labels. We already think about supply chains. We’ve had conversations in this community about fast fashion, about synthetic batting, about the dyes going into waterways.
AI is just the newest material in our creative kit. And like any material — it’s worth knowing what went into making it.
That’s what Thirsty Intelligence is for. Not to scare you away from this new tool. . It can give you the information you need to use AI the way you use everything else: with your eyes open.
Download Thirsty Intelligence for free today. It’s fully cited, genuinely readable, and written for people who care about this stuff. Not just people who work in data centers.