Sustainable AI – are we kidding ourselves?

Chances are, “sustainable AI” isn’t at the top of your daily worry list.

When the team uses ChatGPT, you’re not thinking about polar ice caps or carbon footprints. Most of us don’t.

But with AI creeping into your day-to-day, it’s worth asking: can we ever use it in a truly sustainable way, or are we just kidding ourselves?

Behind every AI chatbot or voice assistant is a data center – a giant warehouse of computers – guzzling electricity and water (for cooling).

In fact, data centers worldwide already consume roughly 460 terawatt-hours of electricity per year (on par with a mid-sized country) and that number is climbing fast.

Training these models also uses enormous energy: one famous neural network consumed enough electricity to power 120 homes for a year and emitted ~552 tons of CO₂ during training.

Using AI isn’t free either. A single AI chat query can draw several times more power than a typical web search. One query is a drop in the bucket, but billions of queries can turn into a flood of demand.

In the rush to adopt the latest shiny tool, we often forget that sometimes a simpler, less energy-intensive solution works fine.

Not every problem needs a machine-learning model. If we’re thinking about the planet, AI should only be used when there’s no easier, lower-energy way to do the job. In other words, if a human expert can do the job just as well, why default to an AI that sucks up 5× the power?

It’s tempting to add AI to everything these days. But chasing trends isn’t always wise (or sustainable). If an AI tool is only a little better than the non-AI alternative (or it’s “cool” but not really needed), it might not be worth the environmental trade-off. Sometimes the greenest kilowatt-hour is the one you never use.

Don’t get us wrong. AI isn’t all doom and gloom for the planet. AI can actually be part of the solution. AI and automation can streamline operations and improve efficiency in ways that ultimately save energy.

For example, AI can optimise supply chains or traffic patterns to cut fuel use, and it helps monitor power grids and emissions for problems.

There’s also the reality that we’re not going to put the AI genie back in the bottle.

These tools bring major benefits, from medical breakthroughs to everyday conveniences, that we won’t give up. The key is making sure the gains truly outweigh the costs.

We should use AI where it delivers big value (especially for sustainability) and think twice where it’s just tech for tech’s sake.

Nothing that eats up energy, water, hardware and human time ever really is sustainable.

Like everything else we rely on (laptops, cloud storage, video calls) it’s not about pretending we can use AI guilt-free. It’s about doing it with our eyes open, and making better choices where we can. Soften the blow by:

Not every task needs the biggest, flashiest model. Smaller, lighter tools often do the job just as well, using a fraction of the energy.

If you’re building something, build lean. If you’re buying something, ask hard questions. Do we need this? Can it run lighter? Is it adding real value?

If you’re in charge of hosting or buying infrastructure, choose providers that use renewable energy.

And if you’re training big models, think about when and where you run them. clean energy availability changes by the hour and by location. A little planning can go a long way.

Carbon offsetting isn’t a free pass, but if you’re going to use AI, balance it out.

Invest in proper, science-backed carbon removal projects (not the cheap “plant a tree, feel better” stuff). It won’t undo the impact completely, but it’s better than pretending it’s not there.

“Sustainable AI” isn’t a hopeless oxymoron. it’s a challenge that demands honesty and intention. No, we’re not going to stop using AI, but we can use it more thoughtfully: employ AI where it truly adds value, and demand cleaner, more efficient tech from providers.

At Telescopic, we take a no-nonsense approach to technology.

We’re not here to shame anyone for using AI. we just want to help you get real value without unnecessary waste. If you’re unsure how to balance AI’s promise with sustainability, it might be time for a tool assessment.

Make sure your AI projects serve your goals and stay true to your values. Progress is great, but progress with a conscience is better.

Book an assessment.