The stories I hear are terrible. Just ONE AI text prompt is equal to running a light bulb for 15 minutes! And look at all the shenanigans we have been doing since the days we discovered AI.
Not a member? Read this article here.
We need it or not, essential or luxury, Ghibli trend or to see myself younger/older, we have been collectively irresponsible in using AI like some kind of cute playground.
It has to END? No. Not necessarily, but the responsibility has to come into the hearts of every user. Just like we have eventually recognized that every paper we print leads to more trees being cut down. This guilt-filled knowledge has to come to us all.
🧠 I Use AI Every Day… and I Feel Kinda Guilty
I'll admit it — I love AI.
As a freelance software developer in Singapore, AI tools have become my digital sidekick. They help me write better code, debug faster, draft client proposals, and even brainstorm article ideas (like this one). What used to take me an hour now takes five minutes. It's productivity magic.
But lately, I've been feeling a little… conflicted. Not because AI is replacing me — but because I'm starting to see the environmental cost of this convenience.
We always talk about AI's impact on jobs, creativity, and ethics — but what about the Earth?
⚡ AI Needs Power — Lots of It
Training an AI model isn't like running a simple app. It takes weeks or even months of constant high-powered GPU computing. One study estimated that training a single large language model can emit over 284 tonnes of carbon dioxide — that's roughly 5 cars' worth of emissions over their lifetimes!
And that's just training. Every time I ask an AI to refactor a JavaScript function or generate copy for a landing page, it has to run on servers somewhere in the world — servers that consume electricity, pump out heat, and guzzle water to stay cool.
So yeah, when I casually type a prompt like "Create a React table with pagination," I'm essentially lighting up a row of data center servers… and I'm not even thinking about it. Oh, the guilt feeling kills me sometimes at night.
🇸🇬 What This Means for Singapore
Singapore isn't a big country. We don't have massive hydro plants or vast solar fields (not enough space, lah). Most of our electricity is still generated using natural gas.
While we are moving toward greener tech, the truth is that many AI services I use are hosted overseas — in countries where data centers rely on fossil fuels.
Ironically, while I'm sipping kopi o' in my air-conditioned HDB flat, a server in the U.S. or Europe is using electricity and water to help me write code faster.
💦 Wait — Water?
Yes, water.
To prevent overheating, many data centers use millions of litres of water for cooling. Some estimates show that a large AI model can consume as much water as a small town during its training phase.
As someone living in a country where we're told not to waste water (remember those "take shorter showers" campaigns?), it feels strange knowing that AI might be doing just the opposite — on a massive scale.
🤖 I Still Use AI — So What Now?
Let me be real: I'm not about to stop using AI completely. It saves time. It helps me do better work. Clients expect faster turnaround, and AI helps me meet those expectations.
But I also think we shouldn't use technology blindly.
I've started asking myself:
"Is this AI query really necessary, or am I just being lazy?" "Am I contributing to a more bloated, energy-hungry digital world?" "Can I find tools that are more environmentally friendly?"
This isn't about going back to the Stone Age. It's about being more mindful.
🧓 A Throwback to Simpler Times
Back when I first started coding (yes, during the dial-up days), we were obsessed with optimization. Every kilobyte mattered. Every line of code had to be tight, efficient, elegant.
Now, we solve problems by throwing more hardware at them. More compute. More cloud. More AI.
Sometimes I wonder: Have we forgotten the beauty of efficient software?
🌱 Is There Such a Thing as "Green AI"?
The good news? Yes.
Researchers and companies are working on making AI more energy-efficient. Some startups focus on training smaller models that run locally or use renewable-powered data centers. Even big players like Google and Microsoft are starting to report on their AI-related emissions.
There's also talk of "AI rationing" — where systems prioritize eco-friendly use cases or limit overuse. Could you imagine a world where your AI tools say, "Eh, maybe don't generate another 20 versions of this same caption, can?"
🧘♂️ My Personal Takeaway
I'm not here to shame anyone (myself included). We're all just trying to make a living, do good work, and survive in a competitive industry.
But we also have to start acknowledging the real cost of convenience.
As developers — even freelancers like me — we have power. We influence how things are built. Maybe we can:
- Push for efficient coding practices again
- Support green tech initiatives
- Be intentional with how often and why we use AI
We don't need to abandon AI. But we can use it wisely.
💬 Final Thought
Every time I use AI, I feel like I'm renting the brain of a digital god… but it's a god with a giant carbon footprint.
Maybe it's time we stop thinking of AI as just magic — and start treating it like the powerful (and energy-hungry) tool it really is.
About the Author
Syed Anees Khan is a Singaporean freelance software developer who develops web sites, web apps, mobile apps and custom bespoke software to exact customer needs. He runs Getcha Solutions, established in 1995, a software development company in Singapore, specializing in custom mobile app development, web development and customized ERP and CRM management systems development.
If you like my writing, now you can subscribe here and be notified of my next one. https://getchacom.medium.com/subscribe