How to use AI without ignoring its environmental impact

AI's possibilities are almost endless. It helps us do more within our work, but are we ignoring its true cost? Before you press 'Enter' on your next prompt, stop and think, "How could I do this better?"
Picture of Ollie Evans

Ollie Evans

Ollie, our Content Executive, makes sure your brand gets noticed. With experience writing content for businesses of all sizes and industries, he knows how to grab readers’ attention and keep them hooked. In his spare time, you can usually find him watching sports, running around Manchester, or cooking up a new recipe in the kitchen.

AI helps us do more.  

It speeds up research and solves whatever tough problem you throw at it.  

But are we ignoring its true cost? 

A single ChatGPT answer uses half a litre of water (Washington Post).  

A single AI-powered data centre can use as much electricity as 100,000 homes (Energy UK). 

Keep reading to learn more about AI’s hidden footprint and how we (and you) can use AI responsibly. 

Can we afford what AI needs?

AI lives in massive data centres. Ones that pull in millions of litres of water to stay cool and burn through incredible amounts of electricity to keep running. 

The more we rely on AI, the bigger those demands get. 

  • By 2030, data centres could more than double their energy use to 945 terawatt-hours, overtaking Japan’s current annual consumption (IEA) 

Water usage is no better.   

  • Training GPT-3 used around 700,000 litres of water (Robotics Tomorrow) 
  • The daily running of just one 100-megawatt data centre can use up to 2 million litres – equal to what 6,500 homes use in a day (Bloomberg) 

Even a single ChatGPT query adds up. It generates about 4.3 grams of CO₂ (Smartly). Now that does sound small, but when billions of queries are made daily, it becomes a serious carbon footprint. 

An image of a row of servers at an AI data centre, all housed in their own container

But your prompt isn't the main problem

Individual prompts do carry a carbon footprint, but this is measured in grams and millilitres.  

The real impact is on the companies that build, train, power and host AI models. 

The coal-powered data centres? The thousand-GPU cluster? Those decisions aren’t made by individual users.  

A single prompt is a drop in the ocean compared to the flood the corporations running the servers are causing. 

Feeling guilty for that ChatGPT prompt or Copilot-powered email only weighs you down and lets the big guns off the hook. 

So don’t beat yourself up. Stay mindful, sure – use AI on purpose, not by habit – but save most of your energy for pushing those companies to clean up and come clean. 

That same mindful habit leads to the next step: knowing when not to use AI. 

Using AI well means knowing when not to use AI​

We’re not anti-AI. We use it for various tasks, such as speeding up research and summarising long, messy notes.  

AI can also be a force for good. It’s already being used to combat climate change and support clean energy innovation (World Economic Forum). 

But just because it works for everything doesn’t mean it’s always the right choice. 

There was a time before AI became mainstream. There’s still plenty to learn if we want to live well with it. This means taking a moment before writing another prompt. Stop and think before you press ‘Enter’.  

We’re not saying to completely ditch AI. But use it with purpose. Progress is still progress, no matter if you took the slower, more considered route. 

How AI is used at MP&Co

As a B Corp, we’re always weighing up the impact of our tools – not just what they help us do, but how much they cost to run. That means looking at the environmental impact of every prompt and shortcut. 

Here’s how we’re trying to do better: 

  • We stop and ask whether AI is really needed. Could a chat or quick Google search do the job? If yes, we do that instead 
     
  • We keep learning about the energy and water use behind the tools we use. The more we know, the better choices we can make 
     
  • We look for offsetting opportunities that match the scale of our use. Things like clean energy initiatives and environmental projects 
     
  • We track our wider digital footprint as part of our B Corp commitments. AI is just one part of a much, much bigger picture 
     
  • We use AI for the right things, like freeing up time to help us do our best work (our top AI tricks are a must-read if you want to do the same) 

We’re not perfect. We’ll probably (definitely) get things wrong along the way. But we’d much rather that happen than stay silent and pretend it’s not our problem. 

As Lucy, our Impact Lead, says: 

“We can and should use AI, but in moderation. It’s a helpful tool that can be programmed ethically and responsibly and used to support us with climate solutions. However, we’re at risk of it being overused and abused.”

What you can do differently

If you are using AI regularly, a few small shifts can make a big difference. 

Here’s how to use AI more responsibly: 

  • Be intentional and thorough with prompts. The more you generate, the more energy you use. Know what you need before you start, and avoid endless re-prompts and tweaks for the sake of it 
     
  • Choose greener tools. Some AI platforms are investing in renewable energy or publishing up-to-date emissions data. Use those when you can. DeepSeek, China’s AI tool, claims to be more environmentally friendly than other AI models (The Eco Experts). Be wary of greenwashing, however. When researching which tool to use, be thorough – some platforms may claim to be environmentally friendly, but how truthful are they being? 
     
  • Offset what you can. If AI is a part of your daily toolkit, find ways to balance its impact – clean energy investment or verified offset projects. Some schemes are better than others, so do your research. Projects led by local communities or focused on long-term protection often have greater impact than quick-fix restoration attempts 
     
  • Chat about it. If you find a new, more efficient way to use AI, tell your colleagues. Share what you’re learning and build awareness among your team – it’s how you build better habits 
     
  • Create an AI policy to get everyone using it similarly. That way, people are deterred from using it too much or incorrectly. It also sets a clear standard for what responsible use looks like across your team 

AI is powerful. But so is every decision you make around using it. It’s about being more mindful of how we use AI and not ignoring its environmental impact. 

This is a learning curve for us all

There’s no perfect system yet. Ditching AI also isn’t the direction we’re telling you to head in. 

The important thing is to take a moment. Be conscious, not careless. Learn the tricks to use AI efficiently and share them around. 

AI isn’t going away any time soon – the way we use it must shift. That means being informed and open about the choices we make and the impact they have. 

This is all still evolving. New lessons will keep coming, and we’ll keep sharing what we find. 

If you’re trying to do the same, let’s swap notes. 

Want more impactful insights like this?

Sign up to our newsletter for real conversations and upcoming events around doing business better (plus plenty on our favourite topics – marketing and growth).

Subscribe now