The problem
AI platforms like ChatGPT have quickly become daily tools.
But how many users know the environmental cost of a single prompt? I started with a provocative idea: What if users saw the water footprint of every AI interaction? Could we help people make more informed, environmentally conscious choices?
The concept
I designed a small modal that reports how much water is used to cool servers after each AI response.
This lightweight UI element delivers a "Did you know?" message in real-time, aiming to nudge awareness without disrupting flow. It’s a tiny intervention with the potential for outsized impact, bringing sustainability into a space where it’s rarely seen.
the research conflict
Initially, I believed that each AI prompt consumed a full 500mL of fresh water, about one bottle.
But deeper dives into research led to surprising contradictions. Some sources claimed each prompt used only 0.3mL. Why the massive gap?
my process
The answer seemed to lie in research methodology.
When I asked ChatGPT, I received low numbers. When I googled manually, I found higher ones sometimes from reputable news sources. This raised more questions than it answered:
Are certain platforms incentivized to minimize perceived impact?
How trustworthy are these environmental stats across sources?
Is there a PR filter in place?
the takeaway
This experiment started with a design question, but it evolved into something messier and more human.
In a world flooded with AI, even data needs to be interrogated. The modal I designed doesn’t just inform, it provokes. It invites users to pause and ask: What else don’t we know about the tools we use every day?
