By Monterey Buchanan
Artificial Intelligence (AY) expert Mustafa Suleman summarized concerns about AI: “The source of the fear is, if you’re building something that’s clever, how do you stop it from out-witting you?” (Johnson 6:24:51-6:24:57). We often picture technology as the thing separating us from the planet we want to protect, not to mention stories like 2001 a Space Odyssey and the rogue learning robot in The Incredibles present apocalyptic visions of a world where technology gets too smart. Artificial intelligence (AI) is everywhere, and can feel overwhelming. Sometimes AI is automatically included in an increasing amount of the technology we use, and many companies and professions are finding ways to use AI in everyday operations. There is also plenty of AI backlash: the risk of misinformation on a massive scale, how and if we should use this technology. Trying to keep up with this complex conversation got me wondering about AI and the climate. All of these perspectives highlight dangers of AI, But how do we figure out what we should fear about AI a climate, and how much we can harness to meet our climate goals? I decided to research and find out.
What is AI?
I had a little previous experience with Chat GPT, Grammarly, and auto-generated email responses, but I decided to go digging for a clearer definition of AI. It turned out there was not just one.
Several sources I looked at had varying definitions of AI. I good starting point came from The UN Environment programme, which defines AI as: “AI is a catch-all term for a group of technologies that can process information and, at least superficially, mimic human thinking.” I found this definition the easiest to get my head around when trying to understand conversations about AI, but the NASA website had several sub-definitions of AI, and included a graphic at the bottom of the page, with a more visual representation of AI
What risks does AI pose for the climate?
One issue that kept coming up in the articles I read was the amount of water used to run the data centers AI relies on For example, in the Associated Press article “Artificial intelligence technology behind ChatGPT was built in Iowa — with a lot of water” by ATT O’BRIEN and HANNAH FINGE Rit describes an upcoming paper on water use and AI: “In a paper due to be published later this year, Ren’s team estimates ChatGPT gulps up 500 milliliters of water (close to what’s in a 16-ounce water bottle) every time you ask it a series of between 5 to 50 prompts or questions…The estimate includes indirect water usage that the companies don’t measure — such as to cool power plants that supply the data centers with electricity.” Increased water consumption like the kind described here will put additional strain on already scarce water resources. In a state like Colorado, which is already drought prone, the increase in AI use could make these issues much worse as AI use increases. The UN Environment Programme article noted similar problems, estimating that “Globally, AI-related infrastructure may soon consume six times more water than Denmark, a country of 6 million,” and also calling out the creation of electronic waste and mining of minerals as negative environmental impacts of AI. AI has the power to exacerbate many of the climate struggles we are already facing.
I t’s not just the AI search tools that have environmental impacts, but the devices we use to access AI in the first place. Phones and computers require mineral mining that can often result in harmful impacts both for the environment and human rights. This is shown in Museum of Modern Art (MoMA) exhibit, “Anatomy of an AI System” by Kate Crawford and Vladan Joler which breaks down all of the resources, labor, and energy needed to create, use and recycle an Amazon Echo Dot. This exhibit walk-through reinforced for me he importance of using the electronics I have for as long as possible rather than upgrading to new technology and requiring additional labor and resource use. It is also a reminder that environmental and human rights are often intertwined. So there are plenty of environmental impacts to be concerned about with AI, but are there any benefits?
Does AI offer any benefits to the climate movement?
There are also some places where AI, if developed and regulated carefully, could be part of the climate solution. The UN Environment Programme notes that AI’s skill at detecting patterns and predicting future outcomes could make it “invaluable for monitoring the environment, and helping governments, businesses and individuals make more planet-friendly choices.” So clearly, there is potential for AI to be part of the climate solution as well.
Efficiency was another area that AI might be able to help with. Several sources mentioned the importance of efficient climate solutions given the ticking clock on the climate crisis. In particular, in her excellent and motivating book What If We Get It Right? Visions of Climate Futures, author Ayana Elizabeth Johnson interviews a wide range of experts about climate-related solutions, including several on how AI might help combat climate change. In the interview “The AI Deluge: Interview With Mustafa Suleman” in What If We Get It Right? Suleman explains: “There’s huge need for efficiency. That’s what I think it’s going to mostly deliver over the next five years. It will make current systems much, much more efficient”(Johnson 6:08:58- 6:09:04). As anyone working on climate issues knows, time is of the essence, so if AI can help humanity respond to the climate crisis more efficiently, it is a tool we may want to consider using.
It is worth noting that AI and technology have already contributed to climate work. I have used a computer and screen reader to help write this article. The 350 Colorado blog keeps activists and newcomers to the movement informed of our progress and how to get involved. (Psst…go to our get involved page!) Johnson even mentions that the interview transcripts for her audiobook are AI generated (Johnson 5:51:53). So, with AI already being used in climate work, and offering potential solutions, but also contributing to the problem, what do we do?
What Do We Do?
As individuals, we can be mindful of when and how much we use AI. Try a google search or other source first, and take time to design specific prompts so that when you do need to use AI, you can get the information you need with fewer questions. But as always, individual actions are not enough. We also need larger society-wide changes to ensure AI will help us address climate change.
Thankfully, several of Johnson’s interviewees have places to start. After emphasizing both individual and legislative responsibility to manage AI responsibly, (Johnson 5:23:41) Paola Antonelli continues: “I think we need to evolve with it, and strengthen our moral and intellectual core so that we can cope with it and push back” (Johnson 5:23:55-5:24:05). It is important to keep ethical considerations in mind, with AI, and Antonelli implies that a paradigm shift among humans as we use AI will be needed.
Finally, Mustafa Suleman suggests a policy change that could be key when charting the “narrow path” (6:32:38) to an AI that helps the climate rather than harming it: “The boring answer is we need competent regulation…we have to have competent, well-paid, well-educated regulators who understand the system, who are deeply embedded in the technical details, who are able to move quickly, make mistakes, get it wrong, change course, no get completely roasted in the media and by companies whenever they do make mistakes. I think we need to give them a lot more love” (6:33:08 -6:33:50). As with so many solutions in the climate movement, much of what will save us comes down to policy changes that may not make a big splash, but are essential if we want a livable climate future.
If you are interested in being part of the climate solution, I recommend volunteering for 350 Colorado.