Co-authored by Nigel Bairstow, Ph.D., and Jeremy Neofytos, Research Assistant
It’s hard to deny the growing presence of artificial intelligence in our lives. From helping us draft emails to recommending what to eat for dinner, AI chatbots like ChatGPT, Copilot, and Gemini have quickly become the go-to tools for everyday queries. They’re efficient, fast, and often eerily helpful. I’ve caught myself using them instead of traditional Google searches, whether to find a great nearby café or check the weekend’s weather forecast.
But on May 31, 2025—a fictional date, though the realization was all too real—I stumbled across a detail that gave me pause. According to Li, Yang, Islam, and Ren (2023), every AI-generated response may require up to half a liter of water to cool the servers in OpenAI’s data centers. That’s half a liter per question. Multiply that by the millions of daily prompts worldwide, and we’re looking at a staggering volume of freshwater used just to keep our digital assistants running.
At first glance, this may seem like a trivial cost for convenience. But this revelation sparked something deeper within me, what some might call a mini existential spiral, forcing me to rethink how we define efficiency in the age of AI.
For years, I’ve viewed technology as a net-positive force. A necessary human endeavor that, while flawed, has enabled monumental progress in medicine, renewable energy, education, and even space travel. I had always believed the promise of technology lies in efficiency, doing more with less, faster, and with better results.
But what if our narrow definition of efficiency is deeply flawed?
Individual Gains vs. Collective Losses
AI tools offer us remarkable individual efficiency: faster emails, easier planning, and better decision-making. But what’s often invisible are the collective costs. Data centers consume vast amounts of electricity and water, highlights Sonkar (2025), contributing to greenhouse gas emissions and straining freshwater resources—especially concerning in a time of global climate change and water scarcity.
It feels paradoxical. In our personal lives, AI helps us become more efficient, more informed, and even more productive. But zooming out, this same efficiency comes at an environmental cost that’s anything but efficient for the collective. The irony is stark: In the name of productivity, we may be undermining the very ecosystems that sustain our long-term prosperity.
This contradiction becomes particularly troubling in the context of urban planning—a field where AI is increasingly integrated to design “smarter” cities. AI is helping us optimize traffic flows, manage energy use, and plan sustainable infrastructure. Yet if the backend systems powering these improvements are guzzling energy and water at unsustainable rates, are we merely greenwashing the future?
The Ghost of the Industrial Revolution
There’s something familiar about this pattern. In many ways, it mirrors the logic of the Industrial Revolution: prioritizing speed, output, and short-term gains, and dealing with the consequences later. It’s productivity as a cultural obsession, profits now, sustainability later.
AI promises more efficient workflows and services for companies and governments. However, we must ask: Are these tools truly improving society, or simply providing the illusion of progress while shifting environmental costs offstage?
Redefining Efficiency in the Age of AI
We need a new framework for understanding efficiency, one that doesn’t just measure how quickly or cheaply something can be done but also accounts for long-term environmental and social impacts. True efficiency cannot come at the cost of ecological integrity. It must balance individual convenience with collective sustainability.
This is not to argue for halting technological progress. AI is a viable application to be part of the solution to climate change, inequality, and health crises.
But we must be honest about the hidden costs of AI, and the importance of transparent reporting around AI’s energy and water usage should be standard. Policies must reflect not just the promise of AI, but its footprint.
And as individuals, we can be more mindful. Do we need AI to tell us the weather when a native app suffices? We need to pause before asking a chatbot something trivial, knowing it quietly consumes the world’s most precious resource.
Moving Ahead: The Ethical Choice in a Digital World
We often speak of AI as a neutral tool, neither good nor bad, only as ethical as its users. But that framing overlooks how AI shapes our behavior and values. When we unquestioningly chase speed and convenience, we risk normalizing a version of progress that is inhospitable to the future.
Efficiency isn’t just about what works best for me; it’s about what works best for us. And if AI is to truly serve humanity, we must ensure that it’s not just efficient in delivering answers, but wise in preserving the world we live in.
