What You Can Do as a Web Builder on Earth Day
One easy thing you can do for the earth is not use “AI” tools.
Consider this as a programmer, web developer, web designer, copywriter, webmaster, etc. The tools include anything branded as generative AI, LLMs, computer vision tools, Copilot, ChatGPT, Bard, Grok, Dall-e, Midjourney, and so on.
If you are also a user who needs to use one of these tools to generate an image description or simplify the reading level of content because the author has created a disabling experience, I am not including you in this broad sweep.
I think Molly White summarizes the current utility of these tools quite well:
[…] they are handy in the same way that it might occasionally be useful to delegate some tasks to an inexperienced and sometimes sloppy intern.
As a bonus, interns are powered by sandwiches. Versus pouring gasoline (almost literally) into the fire consuming our home.
There are plenty of articles demonstrating how ineffective these “tools” are for production-ready code. We know they encode biases into the output. We giggle at people who twist chatbots into giving them deals and screwing over their corporate masters.
But we generally ignore how much damage each of these requests is doing to our planet, our environment, and our future generations. Here is a very brief list of references:
- Training a single AI model can emit as much carbon as five cars in their lifetimes, MIT Technology Review, 6 June 2019
- The carbon impact of artificial intelligence, Nature, 12 August 2020
- Some experts see AI as a tool against climate change. Others say its own carbon footprint could be a problem., CBS News, 26 August 2023
- AI technology behind ChatGPT was built in Iowa — with a lot of water, KCCI Des Moines, 11 September 2023
- The environmental impact of the AI revolution is starting to come into focus, The Verge, 10 October 2023
- AI’s Climate Impact Goes beyond Its Emissions, Scientific American, 7 December 2023
- Managing AI’s Carbon Footprint, Harvard Business Review, 10 January 2024
- Generative AI’s environmental costs are soaring — and mostly secret, Nature, 20 February 2024
- AI likely to increase energy use and accelerate climate misinformation – report, The Guardian, 7 March 2024
- AI may eventually consume a quarter of America’s power by 2030, warns Arm CEO, Tom’s Hardware, 11 April 2024
- What Do Google’s AI Answers Cost the Environment?, Scientific American, 11 June 2024
- Google’s greenhouse gas emissions jump 48% in five years, Ars Technica, 2 July 2024
- AI drives 48% increase in Google emissions, BBC, 3 July 2024
- Planned Google data centre will create almost quarter million tonnes of carbon emissions, The Irish Times, 4 July 2024
- Generative AI is a climate disaster, Paris Marx, 5 July 2024
- AI has a climate problem — but so does all of tech, The Verge, 1 August 2024
- Generative AI is reportedly tripling carbon dioxide emissions from data centers, TechRadar, 9 September 2024
- Here’s What Happens When Big Data Takes Over a Small Town, Sierra Club, 12 September 2024
- AI ‘accelerating’ climate crisis, uses 30 times more energy than search engine: analyst, South China Morning Post, 15 September 2024
- Data center emissions probably 662% higher than big tech claims. Can it keep up the ruse?, The Guardian, 15 September 2024
- Using GPT-4 to generate 100 words consumes up to 3 bottles of water — AI data centers also raise power and water bills for nearby residents, Tom’s Hardware, 19 September 2024
- Microsoft deal would reopen Three Mile Island nuclear plant to power AI, The Washington Post, 20 September 2024
- Unmanaged Climate Risks Undercut AI’s Investment Thesis, John Kostyack at Forbes Sites, 15 October 2024
- Global AI Boom to Triple EU Data Centre Energy Use by 2030, Data Centre Magazine, 27 October 2024
Other things you can do daily that are not explicitly about fake-AI:
- Try on the draft Web Sustainability Guidelines from the Sustainable Web Design Community Group.
- Use the Website Carbon Calculator as a tool to reduce your overall footprint.
- As a proxy to some larger impacts, you can use What Does My Site Cost? to gauge impacts outside of your home country.
Oh yeah, get off the blockchain too. I don’t know a competent, responsible human who is all-in on them, so I doubt I need to make that case to you, dear reader.
While I have your attention, can someone please make that Earth Day site slightly less inaccessible? Wow.
The opening image is my own redrawing of the memed “This is fine” panel from Gunshow by KC Green. The image of the earth is a composite from NASA. In this scene the rest of the planet, particularly the global south, is taking the brunt of the fire while North America is so far unscathed.
3 Comments
Look, if you’ve fallen fully into the AI hate spiral, then I’m sorry for you. From what I understand, the energy consumption of these AI systems can be high, especially during training, but it’s just a fraction of the overall energy usage in the US alone. I’m referring to this Ars Technica article: https://arstechnica.com/ai/2024/06/is-generative-ai-really-going-to-wreak-havoc-on-the-power-grid/
about the true energy cost of artificial intelligence and some predictions about future energy usage. I think it’s worth a read. Besides, I believe that a lot of these criticisms are somewhat irrational, fueled by, I don’t know, maybe too much social media. Just my 2 cents.
Hey Adrian, sorry if my first comment was a bit too offensive. But some of that what I wrote still stands. That I think that overall energy consumption and the statements that are made, the claims, are pretty wild. I think the Ars Technica article provides a good overview of the current situation and many of the claims being made, putting everything into a more balanced perspective.
https://arstechnica.com/ai/2024/06/is-generative-ai-really-going-to-wreak-havoc-on-the-power-grid/
In response to .Mark, responding to this one since it is your later comment but still going to address points from both.
I am not in an
AI hate spiral,considering I acknowledge where it can be useful for people. I am, however, wary of how hard LLMs are being pushed by investors when LLMs are still a combination of stochastic parrots and energy sinks.I appreciate the link to the Ars article from June, written by its senior gaming editor. He is right to challenge assertions in the press. But his claim that other energy usages are worse doesn’t mean it’s ok to add more nor is there evidence the drive for more energy to service genAI will drop. Note that just this week Data Center Magazine (which
connect[s] the world’s largest data centre brands and their most senior executives with the latest trends, industry insight, and influential projects) claims Global AI Boom to Triple EU Data Centre Energy Use by 2030.If you are comfortable with the single Ars article as a refutation of the 24 articles and research papers (so far) I have linked above, then I respect your ability to manage cognitive dissonance.
Leave a Comment or Response