The Environmental Impact of AI
Is Pixar’s Wall-E part of the solution, or part of the problem?
I recently attende the British Psychological Society’s conference on Psychological Contributions to Environmental Sustainability as a guest speaker. I was invited after speaking on AI in Education at Chichester University, where I briefly mentioned the Environmental impact of Artificial Intelligence as part of the quest to reach Artificial General Intelligence. It really was just a single slide on the subject but I did mention I could go on for hours about the hidden cost of this new technology, so I was invited to present at the conference to share my thoughts along with a follow-up Q&A. My presentation, titled “The Environmental Impact of AI”, was more than just a talk on data and carbon footprints, it was a chance to merge my passion for digital innovation with my commitment to ethical, sustainable practice.
I used the Disney Pixar character Wall-E to help illustrate my point throughout. Wall-E is a small, lonely robot left behind on an abandoned, rubbish covered Earth to clean up the mess humans left behind. He is an example of Artificial Intelligence, that over hundreds of years, has developed a personality similar to how traditional AI is moving towards AGI. Wall-E finds a small plant, as intent on nurturing it to prove that Earth is once again habitable, reflecting how AI has been used in farming to monitor conditions, and help conserve our environment. Meanwhile, humans now live on a spaceship and have become lazy and disconnected from the Earth, put their by their own greed and consumerism which has polluted the planet. But what if there was another underlying reason for the contamination anmd destruction of Earth’s ecosystems? The creation of AI, such as little Wall-E, is harmful to the environment in many ways, and so could Wall-E’s mission to help save Earth, come from a guilt that his creation contributed to the destruction? Wall-E was released in 2008, before OpenAI and other corporations had started their frenzied evolution but it’s an interesting new way of looking at the film.
In my talk, I explored the hidden costs behind our rapidly advancing digital tools. We’re used to thinking about AI in terms of its transformative potential,whether it’s shaping education or helping us navigate everyday tasks, but as I pointed out, behind every ChatGPT prompt and facial recognition algorithm lies a significant environmental footprint. From the vast energy demands of hyperscale data centres to the ethically fraught extraction of rare minerals like cobalt and lithium, AI development has a cost we often fail to see.
One of the most striking facts I shared was that training a single large language model, like GPT-4, can consume more water than 500ml for every 5 to 50 prompts. When you multiply that by the millions of interactions these systems process each day, it’s a sobering reminder of the water and energy burden we’re placing on our planet. With the rising popularity of AI image generators which use even more energy, this cost is rising. These are not abstract statistics, they’re tangible challenges, particularly when we consider global inequalities and climate change.
One of the most popular features, was a game of “The Price is Right” but with carbon emissions that had the participants guessing which was higher or lower, training an AI model or Taylor Swift’s jet?
Carbon Emissions Price is Right - the answers may surprise you
The human cost was a particularly emotive section of the presentation. I highlighted how the demand for rare earth elements has led to environmental degradation and social injustices in regions like the Democratic Republic of Congo and the Amazon, often at the expense of Indigenous communities. As someone who has spent much of their career advocating for ethical practice this hit close to home. It was this article in Time magazine about the shocking mistreatment of Kenyan workers whilst training Large Language Models that had me cancelling my subscriptions to several AI tools: https://time.com/6247678/openai-chatgpt-kenya-workers/ The conversation around sustainability in tech cannot be separated from issues of equity, justice, and global responsibility.
But the session wasn’t all doom and gloom. I also spotlighted AI’s potential to drive positive change. From smart irrigation systems in agriculture to biosensors that monitor pollution, AI is also a powerful tool in the fight against environmental degradation. The key, I argued, is to use it responsibly, aligning innovation with ethical standards.
The audience response was overwhelmingly positive. Colleagues appreciated the interdisciplinary approach and the balance of critique and hope. Many commented on how the session pushed them to think differently about the digital tools they use every day, and to use them responsibly. It was a proud moment to see my work resonate with professionals from outside my immediate field.
This presentation also marked a personal milestone. It represented the convergence of so many areas of my work; my advocacy for ethical practice, my ongoing research into AI, and my deep commitment to sustainability and social justice. It was a reminder of why I do what I do and how powerful it can be when technology, and ethics come together.
As I left the conference, I felt inspired not just to continue this work, but to expand it. There’s an urgent need to bring these conversations into more classrooms, more lecture theatres, and more boardrooms. Because if we’re going to harness the power of AI, we must do so with open eyes, open hearts, and a clear sense of responsibility.
