AI and the environment, a long story short

The world is being swept up by AI.

Tools and services are now creeping in to all aspects of everyday life. Whilst this isn’t a ‘new’ technology, the rate at which it is becoming extensively available to us all in the developed world is unprecedented.

Please note: No AI was harmed in the creation of this article.

As a digital agency, we have to be at the forefront of understanding these tools and services, ensuring we can add value for our clients as opportunities arise. That said, our raison d’être means that we have a responsibility to ensure that we (and our clients) are using AI effectively.

A significant part of this is educating our clients, and all audiences, about the impact of AI on the environment. We don’t want to avoid it, we don’t want to tell others to ignore it, we have to be realistic; it’s here to stay and we need to embrace that. It’s also here as a technological advancement, to make a difference. We as humans now have to ensure that that difference is a positive one. Personally, I see huge potential for AI to have a positive impact on our society, but that’s a tangential topic for another day.

AI’s impact on the environment

Bringing things back to the environment; AI has a significant impact due to the energy consumption required to power the data centres behind the technology. Not only are they huge consumers of power (electronic waste), they are also huge consumers of water. On top of these slightly obvious factors, there is also an often overlooked variable to add in which is the rare elements that are mined (often destructively) to produce the supercomputers behind the technology.

Some initial estimates being drawn show that training an AI model (such at ChatGPT) consumes approximately 8 – 10 times the computing power of normal computing activities, such as a Google Search. Once the model has been trained, there is then the ongoing impact of this on the energy requirements to power the deployment of this model to our society.

Anyway, the point is clear and this could be discussed for hours and hours. AI will no doubt use more energy, water and materials than our current technology setups. And may I add, I’m not even touching the ethics of AI in this article!

However, I wanted to lead with the bad news, because there really is good news too. And, there is already so much attention and progress on this – with tools like Green GPT leading the way towards an eco-friendly ChatGPT.

Is it all doom and gloom?

No, it really isn’t, I don’t think.

Peeling back the layers of how AI will become such a part of our lives, even without us knowing so, we’re able to dive deeper to understand the positive effects that AI will have on our planet. It’s important throughout all of this though to consider whether the evolution of AI will lead to significant environmental impacts in a positive manner. Of course, if you just look at AI and its energy consumption directly, it will be seen as detrimental to the environment. But look and think further and AI may well lead to us being able to take significant advances in solving our climate crisis.

Summed up nicely by Golestan (Sally) Radwan, Chief Digital Officer of the United Nations Environment Programme (UNEP), they say:

“We need to make sure the net effect of AI on the planet is positive before we deploy the technology at scale.”

Now that is something I can really get behind. AI is far more powerful than we will ever be at working with data, identifying patterns and predicting behaviours based on previous learnings. By understanding how machine learning (ML) actually works, we’re able to see the potential behind how this could all help identify issues, and solve them. For example, the xView2 project which uses AI to assess disaster hit areas using satellite imagery, allowing responders to know where and how to focus their efforts in delivering immediate aid and assistance to local communities.

There are also advances in AI helping to drive down the energy and material impact that it itself is having on the planet. This shows the potential that AI might well be able to help with stepping us in the right direction towards reaching that net positive effect of the technology.

Again, you can go on and on.

What does this all mean?

In summary, nobody can be sure of the net impact of AI on our planet but we need to play our part in ensuring it is a positive one. We want to work with our clients to ensure they are staying at the front of this curve.

As a consequence, you have to build in real processes to ensure that people are considering the impact of the technology on our planet.

1. Will the use of technology add real value, and can it be justified?
2. Is there anything that can be done to reduce or offset the use of this technology, without losing a competitive advantage?

We’re all excited by AI, and it’s also natural to be unsure of AI! But, let’s all take the first step in educating and understanding that this will have an impact on our planet. That’s a great first step, as it empowers us all to make decisions on how we can all best utilise the technology… for the greater good!