I’m sharing this post here, because I’ve found this diagram useful in clarifying recent discussions about green IT, or sustainability in the tech sector recently.
Here’s the diagram I’ve been using in a few talks and to support discussions.
Sustainability via tech / sustainability via IT
These are the sustainability benefits you can acheive by using a solution to a problem that uses technology. Most of the time, when people talk about tech for good, or sustainability in tech, they’re talking about this. It’s easy to sell solutions, because you’re selling what it does, and you benefit directly, so you’re more motivated to talk about it.
If you sell a solution that makes it easier to use bikes instead of cars, you’ll talk about how you saved emissions by shifting modes of transport to a greener option, or in general using greener ways to meet more or less the same initial need.
Sustainablity of tech / Sustainability of IT
If you’re talking about this, you’re often talking about the direct environmental impacts of using the technology. If we use the example above, we’d be talking about about the energy used to run servers, or make a smartphone used to find a nearby bike to hire and so on.
Here, you’re not really selling what it does, and you’re incentivised to not talk about it, because it can be seen as a reason to not use the solution you’re selling, or otherwise championing.
For this reason, it’s much, much harder to find useful numbers, and companies for the most part do not disclose this in a way that allows you to make meaningful trade-offs.
Are these in tension with each other?
It’s tempting to mentally put these on a single scale, in direct opposition with each other, and hand-wave these away, saying:
If there are positive impacts and negative environmental impacts of deploying technology, then the environmental costs are an acceptable price to pay for the benefits they offer.
I think it’s more helpful and responsible to realise they are separate, but related things. I’m gonna use a bit from a recent report I worked on:
the existence of an upside in any intervention, AI or otherwise, doesn’t automatically free us from the responsibility of thinking about how to mitigate the downsides. When we build systems that store sensitive data, like financial or health records, the fact that they’re useful doesn’t free us from having to think about how to keep the systems safe.
So, think of these more like a minimum bar you would look for, like a minimum level of security you expect when using a bank, or a minimum level of accessibility when visiting a website.
This is why the diagram, instead of being a scale in one dimension, is a two-by-two.
You need both – so if you’re working on things that deliver a sustainably outcome, but not really paying attention to the direct impact of the technology stack you want to be moving to the right into the green quadrant. This is often easier than the other, one, if you’re using sustainable tech to deliver more unsustainable outcomes. In that case, you want to be moving up, into the green quadrant instead. You’re often changing how your organisation works, and you need to be in a position where you are empowered to do that. By definition, most people employed into a role are not hired to do that.
Where to read more
The Environmental Impact of AI report from the Green Web Foundation has more about this argument. Here’s the link to the corresponding part of the document.
This deck goes into a some more detail, about the direct and indirect impacts of tech, both positive and negative. If I was to turn this into a longer post, I’d pull more from it. It cited a bunch of papers that influenced me when synthesising the research to come up with
This deck here I did for Wetransfer a while back is when I first used the two by two in the form in this post.
This post I wrote in 2018 also covers this topic. This one has the graphic that Ive found helps unblock discussions when people are talking about sustainability in a digital context, without being too specific about which one of the two they’re talking about more, because frequently people are having two separate conversations, but using the same words, leading to arguments and misunderstandings.
Two-by-twos have their issues, but in this case, I’ve found them helpful, and they might be useful for you too.