Recently, I had the privilege to speak on a panel of technical experts at the Illinois Innovation Network Conference at Southern Illinois University Edwardsville. Our panel was asked one question:
From your perspective, whether it is human-centered AI concerns, workforce development, ethical decision making, environmental consciousness, or social justice concerns, what practical guidance and applications are critical to your approach for responsible uses of AI?
As an experienced engineer, environmental justice advocate, and environmental researcher, my answer was simple: there is no responsible use of artificial intelligence (AI) in our current societal paradigm. AI exists solely as a productivity and efficiency tool in an infinitely expanding capitalist society. Without limits on growth and resource usage, AI serves as a mechanism to expand production capacity, resource extraction, labor, and capital rather than optimize systems to reduce myriad negative impacts to the environment, human health, and social justice.
AI is marketed and used ubiquitously to maximize production and increase capital. Expanding AI use leads to more labor exploitation, making your job more efficient but giving you more work. While AI can be used to optimize systems to minimize negative impacts, that net decrease is filled with more production.
It doesn’t have to be this way. If the technology belonged to the people—all of the people—then we could collectively decide how it is used and ensure it serves the public rather than billionaires.
AI companies try to appeal to low-income communities because of cheaper real estate. They promise job opportunities and tax revenue, but often most of those jobs are temporary construction jobs and tax revenue is stymied by significant tax breaks granted to companies by local governments eager for the business. Data center establishment also creates housing shortages during construction phases, leading to increased property values that can force residents out onto the street.
While AI growth is a direct result of decades of publicly funded, open-source knowledge, it is now owned and controlled by private corporations. Publicly funded, open-source research, while not necessarily always ethical, is more transparent than current advancements locked behind corporate proprietary paywalls.
AI + finite ecology
The hardware powering this AI boom is housed in data centers, which tend to be sited in resource-strained ecosystems and rural, low-income communities with little power to fight back. Companies take advantage of lower property values. These communities do not have control or input in the development and operation of these technological behemoths built in their backyards.
About 5,000 data centers currently operate in the United States, with that number growing exponentially each year, and our utility infrastructure is not built to handle these new users coming online. More than 150 data centers currently operate in northern Illinois alone. These data centers individually demand enough electricity to power 100,000 homes with newer builds demanding much more energy. This colossal increase in energy demand is requiring utility companies to upgrade infrastructure, and those expenses are passed along to customers, making residential energy bills more expensive.
Modern data centers also use between 300,000 to 5 million gallons of water a day. While the sheer amount of water used by data centers directly for cooling and indirectly for heat and electricity generation is enough of a problem on its own, these facilities are often sited in water-stressed regions. Utilities often divert water to data centers, leaving residential homeowners and renters with induced drought conditions beyond their control. Most data centers require potable water for their cooling systems, diverting millions of gallons of drinking water a day away from residents and to the local wastewater treatment plant. This influx can often exceed plant capacity, presenting a litany of infrastructural challenges for local governments and communities.
Wastewater from data center construction and operation can discharge at elevated temperatures and contain many new or unexpected pollutants like an uncommon and harmful bacterium found in Meta’s wastewater discharge in Wyoming this past summer. Discharging heated water into natural waterways can significantly harm wildlife. Other possible pollutants include heavy metals, PFAS (“forever chemicals”), and antibiotic-resistant bacteria. This contaminated water results from cooling, cleaning pipes, operating “closed loop” cooling systems that use coolant chemicals, and much more.
The proliferation of AI presents additional environmental and social justice issues including air pollution. Data center development has led to increased fossil fuel usage, increased local pollution, and overall growing carbon emissions. While many data centers promise to use renewable energy to power their facilities, we don’t currently have enough renewable energy investment or capacity to meet their growth. The lack of government regulation and permitting enforcement results in AI companies polluting recklessly. Elon Musk’s xAI data center in Memphis runs 30 natural gas generators onsite, creating a startling amount of particulate matter air pollution in an already-pollution-stressed community. A 2025 study of over 2,000 data centers showed that the most common communities sited one mile away from a data center are communities of color. These communities already face constant environmental justice challenges, now exacerbated by the AI boom.
As I’ll discuss in a future blog post, operating data centers on 100% renewable and clean energy would not completely solve this pollution problem. Renewable energy technology often shifts the extraction and pollution burden to rural communities as well as beyond U.S. borders, where most Americans can easily ignore it.
Ultimately, pushing back against water consumption and pollution, electricity and power consumption, and air pollution relies on transparent data from data center construction, operation, and maintenance. Major companies building data centers around the world subscribe communities to non-disclosure agreements, keeping their water and energy data private. These companies are often selective about what data they make available, if any, preventing researchers from compiling comprehensive data sets. Keeping data “confidential” when those data directly impact public health is a danger to communities and ecosystems. Often when these data are made available, they are presented using metrics that are not comparable between different facilities, limiting possible analyses.
The only path forward involves shifting to a degrowth mindset
For technological innovations like AI to truly benefit all of society, we have to collectively shift from resource exploitation to equitable wealth distribution, minimizing resource intensity. Rather than manufacturing scarcity, we can share in reasonable abundance.
Degrowth is defined by the Degrowth Institute as “intentional downscaling of the global economy for the purpose of achieving ecological sustainability and social justice.” This idea contradicts the very core of capitalism: infinite growth. We’ve been taught throughout our lives that if the economy isn’t “growing,” then we are not prospering, we’re struggling. But what if the economy simply met people’s needs instead of redefining them? Degrowth proposes that we work together to exist within the means of our planet. We use and produce what we need for every person to have a good quality of life. That’s it. We don’t mine every last piece of cobalt from the Congo so Americans can have a new iPhone every Christmas. We don’t drill for every last drop of oil across the global south so that the billionaire class can fly their private jets to lunch. We just meet people’s needs.
The concept of degrowth manifests in public utilities via a shift from private for-profit ownership to public not-for-profit ownership. Utilities in the United States are not truly publicly owned. Most utility services like electricity, water, heat, and waste are outsourced to private companies whose sole goal is to generate profits for shareholders. Historically, many services from transportation to telecommunication have been governed as public utilities, but the American idea of a public utility is still a component of capitalist society—in theory, the government places restrictions on monopolies, surge pricing, rate hikes, and more, but each user is still paying a bill to a private company invested in its own success.
Shifting from for-profit private company utilities to not-for-profit publicly owned utilities can give the power back to the people. A truly publicly owned utility is owned by all people in the community and all of them have a say. There can be a democratic governing body within the government that makes certain determinations about utility distribution, but these figures are actually held accountable by the public (a.k.a. the owners of the utility). There is no profit generation, just equitable resource distribution to meet people’s needs and improve everyone’s overall quality of life.
It sounds like a fairy tale, but it’s possible. There are efforts in the U.S. to make electric companies truly publicly owned, advocated by members of the American Public Power Association (APPA), a non-profit trade association. This industry group connects several utilities that have transitioned away from private ownership across the country and provides resources to communities. It’s definitely a step in the right direction.
We should govern AI as a not-for-profit publicly owned utility. This process wouldn’t parallel any existing economic structure within our capitalist society. Instead, this is a complete transition in ownership (not a “buy out”) from private billionaires and shareholders to the entire public community—the working class. If the public truly controlled AI, its development, and its applications, collective decision-making would be for the common good rather than for a few wealthy people’s pockets.
We can transition away from profit-driven applications that result in overall net-negatives to society like surveillance, fossil fuel extraction, weapons development, and war profiteering. Instead, we can develop community-owned management for true democratization and equitable impacts with policy protections. Community governance would protect public health and prevent environmental catastrophes that seem inevitable within our current system of merciless growth that serves only capital. We could protect water, air, and livelihoods.
AI is powerful, but not inevitable—that is just what billionaires want us to think so we follow their lead without pushing back. In reality, it should be our collective decision-making that determines the technology’s future.
Read this and follow me on Substack!