I spent last week at Glastonbury! It was my first time and I really enjoyed myself – great music, fab gluten free food options, camping. What’s not to love?
Of course, I’m a total nerd, and one of the things I was most excited to attend on the first full day was a talk with Professor Gordon Blair about Relationships Between Technology and Nature.
Since attending the Design for Planet festival in October 2022 (which I highly recommend, and intend on attending again this year) I’ve been really interested in the impact that digital services has on the environment. I’ve been looking for time to kick off a more in-depth bit of research on this subject, so I was pretty excited for another opportunity to pop up so unexpectedly!
About Prof. Gordon Blair
Professor Gordon Blair is a Distinguished Professor of Distributed Systems in the School of Computing and Communications at Lancaster University. He started in Computer Science before moving his focus to Environmentalism. He was very personable and funny. You can read all about his work – I wouldn’t want to contribute to the digital litter by regurgitating what is already best said there!
Some quick facts about digital carbon emissions
Did you know that 5 minutes of Googling emits the same amount of carbon as boiling a kettle? You can see how that can add up to a massive carbon footprint in a short while.
In fact, tech emissions are now higher than emissions from the airline industry.
It’s difficult to tell the impact of carbon emissions over all, because big businesses simply aren’t transparent about it. There are a lot of ways to hide how much an organisation is using, and of course many of them find it in their best interests to not disclose. Without this transparency, it makes it extra hard to find out what we need to do to offset the damage.
Finally, and this one blew my mind, 90% of the data created has only been created in the past two years. But the real kicker is that the majority of that data isn’t even used. It’s just stored somewhere, in case someone wants to use it one day.
I’m guilty of this myself – I have emails saved from 2005, a byproduct of Google’s “unlimited” inbox when they launched Gmail. It got me into the habit of keeping all of my emails, just in case one has something useful in it.
Spoiler: most of them don’t. Many of them are probably newsletters I subscribed to and then never read.
Problems with AI
This could be a blog post in itself, and hopefully will be one day, but again I will focus just on what the talk covered.
Of course, we’re all having a great time playing with the new AI tools like ChatGPT. But AI creates a tonne of carbon emissions. Feeding information into AI tools takes a long time, and every minute uses power. While ChatGPT currently uses information that is approx. 2 years old, as use of AI continues, the expectation of more current accurate data will increase, meaning machines will need to be constantly fed more information.
Beyond the environmental impact, Prof. Blair highlighted the issue with bias being perpetuated through AI. The best example he gave of this was when it comes to hiring. If you took the data for most employees in the tech industry, you would likely come up with the expected profile – middle class cisgender white men. So what happens if we teach a machine “this is what a tech employee looks like” and use it when hiring? We end up with more of the same.
What can we do to minimise the impact our work has on the environment?
I plan to continue my research on this topic and hopefully share more in-depth thoughts, but it more or less comes down to flagging the impact at the early design stage.
Prof. Blair refers to “responsible innovation” when speaking of how we obviously need technology to progress, but need to weigh it up with the impact on sustainability – not just environmental, but in all aspects.
Environmental impact should be integrated within the Agile process. There should be a stage at every point where we calculate the impact what we’re trying to create will have on the planet, and find ways to reduce it. We should have sustainability champions within our teams, the same way we have accessibility champions – and the whole team should also be involved in this.
We need to examine our current processes and flag where there may be issues. For example, in Gov design, we use the One Thing Per Page methodology. But is having user journeys that require a new page load for every little step sustainable? Is storing extra data because we can’t have 2 questions on one page having an impact on the environment? I’d love to hear your thoughts!