Artificial Intelligence will Clash with Environmental Justice

Artificial Intelligence systems are scaling worldwide, bringing a series of socio-political, economic, and regulatory challenges. So far, these challenges have been viewed mainly from the perspective of democracy and human rights implications, but 2025 will herald an irreversible shift. These challenges will increasingly intersect with demands for environmental justice that rebel against the socio-environmental impacts of AI data centers.

The AI supply chain relies on creating a globally deployed infrastructure, which has driven the extraction of more minerals, energy, and water in various territories. Data centers have been particularly contentious due to, among other reasons, their rampant consumption of electricity (which jeopardizes countries' climate goals) and freshwater (which could lead to disruptions in the global supply chain), as well as its e-waste generation. Some say generative AI adoption could generate between 1.2 and 5 million metric tons of e-waste in total between now and 2030.

These potential catastrophes have catalyzed growing resistance to the AI boom, led by diverse communities. Groups in the United States, Latin America, and Europe have opposed the AI-led demand for fresh water in drought-stricken areas. Citizens in Ireland have raised concerns about the carbon footprint of data centers within its borders. The problem is even more pronounced in countries like South Africa, where electricity shortages have been a national scandal in recent years, even without AI taxing the grid. Using land and clean energy to deploy these private infrastructures has also been controversial in places as varied as the Netherlands and Singapore.

I study the resistance and responses of territories in Latin America facing sustainability concerns over data centers. And I know 2025 will be a special year: 2024 legislation to temper AI's environmental impact in the US and the EU can influence the industry and legislation worldwide. In turn, technology companies and technologists are looking to introduce more sustainable practices in data centers. Governments, even in the Global South, are advancing plans to attract AI investment that include sustainability concerns – though not without controversy. Moreover, if we can trust the forecastse regarding how data centers will increasingly host advanced AI workloads, we can expect that socio-environmental conflicts will increase, especially in Latin America, too.

2025 is thus a key moment to design and deploy better, transparent, and participatory public policies and measures. I will introduce a publicly available tool that systematizes and organizes data on the socio-environmental impacts of AI, the diverse demands it is placing on land, and the successful and unsuccessful responses of companies and local and national governments. I’m particularly interested in the conditions faced by Latin America concerning its position in the supply chain and the unequal power relations in a global scenario that directly impacts the power of territorial communities.

This work is not a matter of providing recipes or silver bullets for a problem at the heart of 21st-century capitalism. Instead, it is about examining the complexities of digital sustainability amid an ecological and climate crisis, systematizing significant information, increasing its transparency to stakeholders, and advancing responses worthy of the people in affected territories and communities, especially in the Global South.