When we think about AI, it's easy to overlook the material aspect of this rapidly growing technology. Behind our chatbots and image generators lies a vast infrastructure of data centers and servers, consuming huge amounts of water and electricity. We're quickly learning how much these data centers are consuming, but we're not talking enough about what they're putting back into our air. You guessed it—pollution.
The Human (and Financial) Cost of Invisible Emissions
New research from the University of California, Riverside (UCR) and Caltech shows that pollution from AI could rival or even exceed emissions from entire industries, potentially surpassing all vehicle traffic in California or even coal-based steelmaking. Researchers are also predicting that by 2030, air pollution from AI could cause up to 1,300 more premature deaths a year in the U.S. This is more than just a warning—the challenges of AI sustainability are very real.
Companies developing AI are not yet required to report on the impact they have on air quality, thus giving the general public little visibility into the harm. They may share sustainability reports with carbon emissions and water usage, but they're not sharing their output of air pollutants, which can cause health issues like asthma, heart attack, stroke, lung cancer and cognitive decline. Letting data center emissions go unchecked won't just impact our lives physically—it will also cost us financially.
The health costs associated with AI pollution are projected to hit $20 billion annually. Whether it's missing school or work due to sickness or constant treatment of respiratory illness, heart disease and other pollution-linked conditions, this all gets added to the tab—and the public picks it up.
Putting Marginalized Communities at Greater Risk
Because data centers are often located in rural, low-income areas to save on costs, these communities will be forced to take on more of the burden. People in marginalized communities already face higher rates of asthma, reduced access to healthcare and disproportionate exposure to environmental hazards, but now they will bear the brunt of industrial emissions tied to AI's infrastructure. These data centers are doing more than just polluting our air—they’re adding to our ongoing public health crisis and social injustice.
The Future is Not Set
Aside from giving Terminator 2 a rewatch, we don't have to rely on John Connor to be our hero—we need to demand more from tech companies. The researchers behind the UCR and Caltech studies have solutions in mind, like transparent reporting and the development of “health-informed AI” practices. Just as we’ve demanded corporate climate pledges and carbon accounting, the time has come to demand pollution accountability from big tech, factoring in the true human cost of digital innovation.
We can also push for cleaner power grids, stricter emissions standards and more efficient data centers, while asking policymakers to mandate transparent reporting and demand compensation for the communities most impacted by industrial emissions.
The Bottom Line
We're still in the infancy of AI development, but the hidden costs of artificial intelligence are becoming too real to ignore. The tech that has the potential to change the world shouldn’t be creating a health crisis in the process. You don't have to be anti-AI, but we should all be pro-responsibility.
Sources for All:
-
https://www.caltech.edu/about/news/air-pollution-and-the-public-health-costs-of-ai
-
https://news.ucr.edu/articles/2024/12/09/ais-deadly-air-pollution-toll?utm_source=ADSA&utm_campaign=3b1627ea04-EMAIL_CAMPAIGN_306_COPY_01&utm_medium=email&utm_term=0_5401c7226a-3b1627ea04-481332474
-
https://molekule.com/blogs/all/indoor-air-pollution-unhealthy-ingredients?_pos=2&_sid=14c3b37de&_ss=r
- https://hsph.harvard.edu/news/office-air-quality-may-affect-employees-cognition-productivity/
