Power-hungry prompting and dirty datacentres: The AI dichotomy for a planet-friendly business

“Not another AI blog post!”, I hear our Alex and many others cry. Bear with me though, as this one’s a little different. Instead of practical tips around how to craft the perfect prompts or the magic of “vibe coding”, I’ll be sharing some thoughts about how the use of AI causes a moral and ethical dilemma for a business that’s trying to work with brands that put the planet first.

The dilemma? What you may or may not be aware of is that AI, for all its lauded benefits, has a hidden cost. And that is the cost to the environment. 

Remember those cutesy plastic-wrapped effigies of ourselves that were doing the rounds on social media a few months ago? In order to create them, tonnes of carbon will have been belched out into the atmosphere, and hundreds of thousands of litres of freshwater will have been used to keep the data centres cool.

So concerned was OpenAI’s CEO, Sam Altman, at the time, that he had to restrict the use of ChatGPT for this purpose:

Hint: “melting” is the operative word here.

Moreover, there are some serious privacy concerns raised about the trend as well, which may be new to you.

“The cloud” is still a physical thing

This is just one small example of how digital has hidden costs, and they need to be measured. I’ve written at length in the past about how all websites need to be much more sustainable and climate-friendly, and thankfully, I’m not the only one who thinks that.

So you may be forgiven for thinking at this point that I’m going to advocate that we all abandon AI as quickly as possible to stop the world burning and to feel better about ourselves. 

However, that’s not what I’m going to do. Because I’m acutely aware that AI is here, and here to stay. I’m also aware that it's proving quite tricky to measure the actual environmental impact of AI use, so, given that I’m not a scientist, I won’t overstep the mark here. Finally - and as I talk about a little later - things perhaps aren’t as bad as we think they are, and we have to consider productivity and efficiency benefits as part of this equation too.

What does the literature say?

Instead of putting my virtual flag in the virtual sand, I feel this blog post should become a signpost for further reading on the topic, should you wish to explore this issue further. Here are just a few pieces to help you become more worried, or put your mind at ease…

NB. The parts in bold have been emphasised by me.

First up is a sobering piece from the Massachusetts Institute of Technology (MIT), which dives incredibly deeply into the topic. The tone of the article is negative regarding the environmental risks of AI. It frames the rapid growth of AI as a serious ecological challenge, using strong language like "staggering in scope" and "alarming" to describe its environmental impact.

“If you’ve seen a few charts estimating the energy impact of putting a question to an AI model, you might think it’s like measuring a car’s fuel economy or a dishwasher’s energy rating: a knowable value with a shared methodology for calculating it. You’d be wrong.

In reality, the type and size of the model, the type of output you’re generating, and countless variables beyond your control - like which energy grid is connected to the data center your request is sent to and what time of day it’s processed - can make one query thousands of times more energy-intensive and emissions-producing than another.”

https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech

***

Next up is Andy Masley’s Substack post. Again, it’s lengthy, and Andy goes into a very impartial analysis of the data, but his main tenet is that we are potentially missing the point here with a misplaced focus on personal AI use. The author uses a "cheat sheet" format to debunk popular environmental objections. So while the tone of the piece is not explicitly in favour of using AI, it points out that we have much more effective opportunities to reduce our environmental impact, like reducing flights and switching to an electric car (or even ditching a car altogether!), than ditching AI.

“If everyone in the world stopped using ChatGPT, this would save around 3GWh per day. If everyone in the world who owns a microwave committed to using their microwaves for 10 fewer seconds every day, this would also save around 3GWh per day.

Imagine you met someone campaigning to convince everyone to decrease their microwave usage by 10 seconds per day. Your first reaction would be “There must be much more effective things to campaign for that would have a lot more climate impact. This must be really small compared to global energy demand.” You’d be correct. Campaigning to stop ChatGPT is exactly like campaigning to use microwaves for 10 fewer seconds.”

https://andymasley.substack.com/p/a-cheat-sheet-for-conversations-about

***

Which leads us nicely onto the final piece, written by (disclaimer: friend of the show) Lorna Kirman and Rory Mitchell at HRW Healthcare. Their piece reflects where I think I am, mentally, right now with AI. It highlights the negative discourse around the topic, but is much more positive and reassuring about the actual environmental risk posed by an individual's use of AI.

“Pointing fingers at AI tools while empty private jets criss-cross the globe to reposition them for their owners? That’s a distraction. 

Demand more: Holding companies, industries, and governments to account is one of the most powerful actions we can take. 

For example, your pension fund – often overlooked – could be financing fossil fuel giants. One simple switch could do more than a lifetime of prompt avoidance.”

So what do you think, Rob?

As I alluded to above, my thinking is now more in line with that of Lorna and Rory (and Andy). I feel far more comfortable with AI to augment what I do in the workplace and working with our clients (despite existential concerns, too).

I’m turning to it more and more every day to support different aspects of what I do. It’s saving me time, helping me learn more quickly and is opening up my mind to different perspectives and ideas.

So in summary: I need to be aware of other things I do in life - that potentially have a greater impact on the environment - but recognise that I’ve added another digital tool into my toolkit that has an impact, even if, individually, that impact is relatively minor.

As the MIT piece states, though: “...as more of us turn to AI tools, these impacts start to add up. And increasingly, you don’t need to go looking to use AI: It’s being integrated into every corner of our digital lives.

Keep in mind that the ways people use AI today - to write a grocery list or create a surrealist video - are far simpler than the ones we’ll use in the autonomous, agentic future that AI companies are hurling us toward.”

But, and it’s a big but: we need to compare the global impact of AI use to the global, rather than personal impact of other (more environmentally-harmful) things.

In other words, I think I’ll have to write a piece on this in a year’s time when this debate has rolled on!

A quickfire primer on the environmental impact of AI when prompting

Finally, here’s a brief Q&A around AI use for marketing:

Q: Does generating an image use more energy than just a basic text prompt?

A: Surprisingly, no. Because a text prompt may require more parameters to interpret, it can often have a greater energy need than image creation, which often works on fewer parameters.

Q: So does that mean AI-generated videos aren’t quite as power-hungry then?

A: Unfortunately, not. They’re actually the worst culprits; in order to create a five-second clip, it’s estimated that it would use enough energy to run a microwave oven for over an hour. In comparison, a generated photo may only be the equivalent of using the microwave for a few seconds.

Q: Surely modern datacentres are powered by renewable energy, so this doesn’t really matter?

A: Because the energy density of data centres is so high (and the sun doesn’t always shine, and the wind doesn’t always blow), fossil fuels are still a significant energy source for data centres. So much so that big tech companies are proposing new nuclear power plants to power the increasing energy needs for ever-growing data centres.