The real climate impact of AI goes beyond resource consumption

Written by Jack McGovan / Edited by Libby Langhorn

Technology ethicist Paul Schütze first became interested in AI when he came across research around five years ago predicting that the technology would become the latest spearhead of digital capitalism. “I was interested in this most recent development of business as usual,” he said. “What is this wheel that we’re stuck in, and why do we continue to be stuck in it?”

That prediction, or perhaps prophecy if you were to ask a tech enthusiast, has unfortunately come true. ChatGPT and other large language models (LLMs) have since become household names, following in the grimy footsteps of social media algorithms, cryptocurrency, and NFTs. These LLMs are lumped into the broader acronym AI, which includes a range of other technologies like machine learning algorithms used to do climate modelling. 

LLMs in particular have come under heavy criticism since their release, not least because they steal work from artists, and they’re damaging our ability to think critically. By 2027, the water demand for AI is expected to reach up to six times the annual consumption of Denmark. By 2028, AI infrastructure in the USA could use enough electricity to power 22 percent of homes in the country for an entire year.

Schütze, who works for the University of Osnabrück, Germany, has developed a climate critique of AI that moves away from the resource consumption angle as part of his PhD dissertation that he submitted in January. He argues that AI reproduces the values that have gotten society to this point in the first place—extraction, productivity, and efficiency—and more of the same is only going to make things worse. “Even if AI were to be sustainable in the sense that it will be all efficient, consume no resources, and have no environmental footprint, it still would reproduce the very system that caused the climate crisis,” he said.

His underlying argument is that it’s not just enough for us to move away from fossil fuels to solve the climate crisis. We also need to shift the values of our society away from extractivism, productivity, and efficiency, and move instead towards values such as care, compassion, and community. The technologies we choose to use are a reflection of our values, and AI in its current form is only further entrenching the status quo.

X reveals a weakness in left-wing rhetoric

Good principles mean nothing when you materially support a fascist propaganda machine.

SowerJack McGovan

A tangible demonstration of this can be found in cars, said Schütze. Cars are a mode of transport that allows a person to move between different points without having direct contact with other humans. A society that builds that kind of infrastructure has a certain set of values attached to it, like, for example, prioritising individual autonomy over all else. The more people use cars, the more those values become ingrained in society.

On the flipside, a society that builds out a robust public transport system would have different values. That society would value collectivism and community spirit more, as transport would only be possible with direct contact with other humans—shared spaces become more important than individual autonomy. To that end, while electric cars might be more sustainable than fossil-fuel vehicles in terms of emissions, they still promote values that uphold the system that brought the climate crisis in the first place.

That’s not to say that people who use cars or AI are evil; most people just want to live their lives and use the tools that are available to them. It’s just about recognising that technologies are more than how their end users interact with them. AI is not just an app: it’s imbued in the ideology that birthed it, and it’s the infrastructure that was built to support it—the data centres, hardware, and labour. “You can never really talk about ChatGPT, LLMs, or other algorithms without also talking about this other infrastructure,” said Schütze.

Working less is the most underrated climate solution

Fewer work hours can lower emissions. So why isn’t the climate movement fighting for our collective right to relax?

SowerJack McGovan

Rather than trying to bring humanity together to help us deal with existential crises like the climate crisis, AI is further siloing us into individuals, separating us from real human interaction. People are so alienated by modern life that they’re turning to an inanimate piece of software for emotional support. Rather than spending more time with our children and enriching their lives with real human contact, we’re instead putting software in their stuffed toys to interact with them—software that tells them how to find matches and gives detailed explanations of sexual kinks. 

Instead of going along with these developments, why don’t we take a step back as a society and ask whether this is something we actually all want for our future? That, Schütze said, means we need to sit with the reality of our situation and face it head on. Dealing with the climate crisis is going to be an uncomfortable process, it means we’re going to have to give up the comfort and ease that AI promises us. “Although that can make you feel very overwhelmed, I think it’s important to really let reality sink in, and then let it motivate you to act,” he said.

While Schütze doesn’t have a lot of hope that we are going to turn things around in time to address the climate crisis, he doesn’t want to give into nihilism. He advocates instead for people to live authentically in a way that aligns with their values; collectively we can steer the world in a different direction. “I want to live in a world with more solidarity, for instance,” he said. “So even though I don't believe that a better future for all is on the cards right now, I can still build community and live according to my values to try and make that future a reality.”

Find out more about Schütze’s work in his own words by reading a blog post he recently published.