After reviewing the reading materials for this blog prompt, I kept feeling the curious question rising of:
"what is the digital and carbon footprint of artificial intelligence use and of more powerful technology now?"
When delving into Neil Selwyn’s 2021 article Facing up to the Dilemma of Sustainable Digital Futures, I found his exploration of the environmental impacts of digital technologies correlates to the unsustainable carbon footprint with the increase of human reliance. From manufacturing to data storage, Selwyn details how the “disposal of e-waste is another major environmental burden” and one that compiles on the existing burdens of AI usage.
He states strongly how we, as a society and individual people, cannot continue to consume digital technology at current rates due to the foreseen and unforeseen detrimental consequences. AI is so early on in the mass production and availability for general use now that we cannot determine the exact impacts it is making -and will continue to make- on the environment… but we can safely predict that it’s not feasible.
Another important point that Selwyn discussed in his article is the human reliance on AI and digital access now. It is inextricably not slowing down and while the technological advances are revolutionary in theory, the practices could be the eventual determining factor of our downfall. Not to give way to the concern of AI taking over, but with regard to the environmental impacts and carbon footprint, AI may take over as one of the largest mass waste producers.
My problem is how many different sources share conflicting results and opinions. I can’t decipher, as I have not conducted any studies myself, what information is reliable. Social media platforms tend to lean towards hyperbolic claims of the damage that AI use leads to while an instructor for another course at UVic requires AI usage as part of their curriculum. The information I am receiving is inconsistent and it’s hard to believe either side.
For example, in this article, Using ChatGPT is not bad for the environment by Andy Masley (2025), he argues strongly how many everyday activities like using a computer or running the vacuum cleaner use much more energy than using an AI software like ChatGPT. While reading through the article, the researcher does seem well-informed and uses many heavily-researched examples, however I do find the tone of the paper to be so assertive that I almost mistook it for satire initially. I also want to point out that the website Masley’s article is published on is a self-publishing site. This means that it may not go through a reviewing process by an external body. These kinds of things are important to note to determine the motivations and credibility behind published materials.
Alternatively, this journal article by Kate Saenjo and The Conversation US titled, A Computer Scientist Breaks Down Generative AI’s Hefty Carbon Footprint (2023) in Scientific American, tends to gravitate towards the other side of the AI use spectrum and argues that the usage of water and energy specifically is actually incredibly harmful towards the environment. Saenjo compares the creation of an AI model called BERT consumed the equivalent energy to a “round-trip transcontinental flight for one person.” Granted, this is measured for the full creation of this particular model. Saenjo goes on to present other comparisons in energy usage and determines that all of it (AI creation, maintenance, and usage) is more impactful to environmental damages than our planet can sustain.
The image above is how I’m feeling about the argument. Ultimately, as an advocate for environmental protection measures, I tend to lean towards the presumption that the damages to the environment will lead to some catastrophic outcomes. At the very least, we know, as pointed out by Selwyn (2021), that the environmental impacts are unsustainable without the adoption of climate-friendly digital practices.
Leave a Reply
You must be logged in to post a comment.