Can Artificial Intelligence and Environmental Justice Coexist?

The creation of generative AI tools such as ChatGPT, Gemini, and Claude are widely vilified as a crime against the climate. The massive amounts of energy, water, and natural resources used in their creation is well-documented--or at least as well as restrictive interpretations of proprietary knowledge will allow. Given this, it is reasonable to ask if it is possible for AI to serve as a tool for environmental justice.

 

I'm skeptical too.

 

Nonetheless, FRI is working with a team of UCD computer science students to create an AI-based technology to serve environmental justice advocates and communities long subjected to environmental injustice. We are trying to combine the power of LLMs with the expertise of EJ to create a program that can serve community needs. This customized program would help communities to quickly understand and analyze technical, scientific, legal, and government documents from an environmental justice lens.

 

Environmental justice advocates spend long hours analyzing environmental impact reports (EIR), permit requests, research studies, policy pieces, proposed legislation, and other specialized documents--often hundreds of pages long. They do this despite a lack of formal training, long working hours, and other responsibilities such as care-giving. They do this because they care about the health of their families, their children, and their neighbors. They want to have safe places to live, work, and play, where they do not have to worry about the air they breathe or the water they drink shortening their lifespans.

 

Yes, there's an irony in using a technology that relies on resource-intensive data centers to help communities fight environmental harms. Part of this experiment is tracking the compromises we make along the way. Feminist work demands that we account for process. How we get there is just as, if not more important, than where we ended up. The two are inextricably tied together, which is precisely why so many of us hold deep misgivings about the "promises" of AI.

 

Still, I hold out hope for other speculative futures and want to engage in imaginative experimentations to get there. Such experiments are not without risk and are far from innocent. Can we mitigate the environmental massacre that is generative AI by turning this Frankenstein's monster against its creators? Can we use this as a harm reduction tool to prevent the proliferation of more anti-life technologies, projects, and constructions?

 

Paperwork, processes, and policies that are supposed to protect communities from environmental harm can and are put to more nefarious uses. It's an act of hope and speculation to see the extent to which LLMs can be put to better uses.

 

If this sounds like a project that interests you, we could use help. We need stories of times when the "experts" tried to hoodwink communities and were caught. And we need the paperwork--permit requests, EIRs, calls for comment, legal filings, reports... As always, those who came before us offer up examples and inspiration for the future. These examples train environmental justice leaders today, and we will see if they can also train our non-human silicon-chip companion.

 

Do you have a story of environmental injustice overcome or embattled? Can you help us to dig up the bureaucratic dirt on such cases? If so, please reach out and join this exercise in practicing speculative, skeptical hope.