By Kelly Gibbs
Artificial Intelligence (AI) has integrated itself into almost every aspect of society, yet the general population remains mostly unaware of the extent of its prevalence and detriments. As of March 2025, 77% of electronic devices used some form of AI, but device users assumed it to be only one-third.[1] The mass increase in AI use poses a massive and multifaceted threat to the United States—specifically through its endless energy consumption, depletion of freshwater resources, and perpetuation of environmental injustice and public health crises.
AI Energy Consumption and Air Pollution
The energy AI necessitates is far greater than that of the standard internet. A single inquiry into ChatGPT requires ten times the amount of energy as a typical search engine, and a single AI-generated image consumes the same amount of energy as a phone being fully charged.[2] However, it is not the AI platforms that require the energy; it is the physical data centers that host the technology to power AI programs, and the numbers have increased exponentially. The size of the data centers and the quantity of equipment within them is the real energy consumer. In 2023, data centers used over 4% of total U.S. electricity, and that number is expected to jump to 12% over the next three years.[3] ChatGPT alone is estimated to release over 494 billion grams of carbon dioxide annually.[4]
AI Water Consumption
Furthermore, AI requires an exuberant amount of water to run. The substantial amount of hardware and equipment within data centers rapidly generates intense heat. The high heat levels necessitate cooling systems to reduce the heat and keep the technology at optimal temperatures. Large data centers can consume up to five million gallons of water per day, roughly equivalent to the water use of a 10,000-to-50,000-person town, to cool the infrastructure.[5] AI’s increasing usage will only worsen these conditions. By 2028, AI, in the U.S. alone, is projected to require 720 billion gallons of water annually.[6] States throughout the U.S. are already facing the detriments. Texas, Arizona, New Mexico, Utah, and California are all facing water shortages after welcoming data center development.[7] This excessive water consumption depletes the limited amount of drinkable water available on earth, destroys ecosystems, and degrades soil.
AI and Environmental Injustice
The growth and use of AI further exacerbate environmental injustice. Environmental injustice is the inequitable overexposure of low-income, impoverished, and minority communities to environmental hazards and toxic chemicals.[8] Due to the expensive nature of data center technology, companies seek to cut the costs of data center building, maintenance, and operation. One of the most effective methods of doing so is by taking advantage of the low land costs in low-income and minority communities. These communities are advantageous not only because of their low costs, but because data center pollution is more likely to be overlooked if it occurs in low-income, minority communities.
As of April 2025, Big Tech—Google, Meta, Microsoft, Apple, and Amazon—have spent $200 billion building and expanding data centers in the low-income, rural South.[9] A prime example of environmental injustice is Elon Musk’s AI company, xAI. The xAI supercomputer, Colossus, is located in a community where the annual household income is below $25,000, and the cancer rates are four times the national average.[10] Now, since the establishment of the supercomputer, the concentration of nitrogen oxide, an air pollutant linked to several health risks and illnesses, has increased by 79%.[11] Additionally, the government provides incentives for AI structures in communities that have experienced decades of environmental injustice, funneling public subsidies away from community resources and toward AI companies.[12] All helping the industry, but leaving citizens struggling with health issues, pollution, and depleted resources.
Several major pieces of environmental legislation have included provisions to improve environmental justice: Title VI of the Civil Rights Act, the National Environmental Policy Act (NEPA), the Clean Air Act, the Clean Water Act, and the Safe Drinking Water Act all include legal mandates to protect disadvantaged populations.[13] However, these established pieces of legislation, the initiatives, and policies have failed to establish enforceable rights for citizens and regulation across all sectors. Therefore, a new act or policy that has clear environmental justice standards, including strict oversight measures and liability frameworks for the communities impacted, needs to be adopted.
Public Health Detriment
Furthermore, the environmental damage from data centers also increases public health issues. The air pollutants released by data centers are often called “silent killers.”[14] The air pollutants released the most are PM2.5 and nitrogen oxides, which are linked to several health conditions, such as asthma, lung cancer, heart attacks, and death.[15] These pollutants are the second-highest risk factor for noncommunicable diseases and are “non-threshold” chemicals, meaning there is no safe level of exposure.[16] As the number of data centers rises, millions more U.S. citizens are put at risk of life-threatening diseases. This increases the need for government action—air pollution is expected to cost the U.S. over $20 billion in public health burdens by 2030.[17]
The U.S. government has tried and failed to address toxic pollutants. The Clean Air Act (CAA) regulatory gaps and weak oversight measures allowed the issue to be dragged out and underaddressed for decades. A key example of this is the CAA subjecting 188 toxic air pollutants to monitoring, but since the establishment of the CAA, there has been no adaptation to account for new or strengthened versions of toxic air pollutants[18] Adjustment to developments is necessary for public safety, and with the prevalence of AI increasing, it is highly pertinent that its air pollution is included in future legislation. Without action, data center pollution will continue to add pressure and expenses to an already failing healthcare system, costing lives and billions.
Taking Action – Moving Forward
The U.S. needs to take action. More than 190 countries have already adopted ethical recommendations for AI use that address its detriment to the environment.[19] For example, the EU has adopted the EU AI Act, which includes provisions for required reports on energy consumption from high-risk and generalized AI systems throughout both the development and usage stages.[20] The U.S. government can and needs to take inspiration from other nations and put policies in place to address AI over-expansion and its consequences. The current key policy gaps lie in energy regulation, emissions monitoring, and approvals and grants for development locations.
The large AI boom has given technology companies an exponential increase in revenue and investment. Therefore, subsidies or tax credits for renewable energy use provide less of an incentive than strict policies. Strict policies require AI companies to internalize the cost of their environmental damage; this can be achieved through regulatory and market-based tools, such as pollution fees, remediation funds, or liability frameworks for damaged communities. At the rate climate change is increasing, the U.S. does not have the luxury to remain complacent. AI is not going to disappear, but the U.S. government can and must do all that it can to protect the environment and citizens’ health.
[1] Contributors, NU Editorial, and Timothy Prestianni. “131 AI Statistics and Trends for 2026.” National University, March 4, 2025. https://www.nu.edu/blog/ai-statistics-trends/.
[2] Association, National Education. “Environmental Impact of AI.” NEA, June 20, 2025. https://www.nea.org/professional-excellence/student-engagement/tools-tips/environmental-impact-ai.
[3] Association, National Education. “Environmental Impact of AI.” NEA, June 20, 2025. https://www.nea.org/professional-excellence/student-engagement/tools-tips/environmental-impact-ai.
[4] Education, Better Planet. “The Environmental Impacts of Artificial Intelligence.” Better Planet Education. Accessed January 30, 2026. https://betterplaneteducation.org.uk/blog/the-environmental-impacts-of-artificial-intelligence?gad_source=1&gad_campaignid=22836242556&gbraid=0AAAAADmeCSVYdnZa5B6SlkFpXhmW8Sd-S&gclid=Cj0KCQiAp-zLBhDkARIsABcYc6vVBmG4qBKN8dTQkJQO6C2T7M1WLcjmj5JvdC7W1bcIyt5JF_I5VxgaAiITEALw_wcB
[5] Yañez-Barnuevo, Miguel. “Data Centers and Water Consumption.” EESI, June 25, 2025. https://www.eesi.org/articles/view/data-centers-and-water-consumption.
[6] “Updated February 2026 a No Brainer: How Ai’s Energy and Water Footprints.” Food and Water Watch. Accessed February 13, 2026. https://www.foodandwaterwatch.org/wp-content/uploads/2026/02/FSW_2602_AI_Water_Energy_UPDATE.pdf.
[7] Maggie Morales, “5 States Facing Critical Water Shortages as AI Data Centers Multiply: Waterless Co Inc..,” Waterless Co. Inc., January 6, 2026, https://www.waterless.com/blog/-5-states-facing-critical-water-shortages-as-ai-data-centers-multiply.
[8] Landrigan, Philip J, Virginia A Rauh, and Maida P Galvez. “Environmental Justice and the Health of Children.” The Mount Sinai journal of medicine, New York, 2010. https://pmc.ncbi.nlm.nih.gov/articles/PMC6042867/.
[9] Dulani, Jai, Brandon Dill, Sarah van Gelder, Jai Dulani Jai Dulani, Emily M. Bender, Alex Hanna, Jesse Roth, et al. “Big Tech Data Centers Compound Decades of Environmental Racism in the South.” Truthout, September 20, 2025. https://truthout.org/articles/big-tech-data-centers-compound-decades-of-environmental-racism-in-the-south/#:~:text=us%20as%20human.%E2%80%9D-,As%20the%20South%20becomes%20the%20new%20epicenter%20of%20data%20center,fuel%20plants%20in%20the%20area.
[10] Dulani, Jai, Brandon Dill, Sarah van Gelder, et al. “Big Tech Data Centers Compound Decades of Environmental Racism in the South.”
[11] Dulani, Jai, Brandon Dill, Sarah van Gelder, et al. “Big Tech Data Centers Compound Decades of Environmental Racism in the South.”
[12] Mario DeSean Booker. “Digital Redlining: Ai Infrastructure and Environmental Racism in Contemporary America.” World Journal of Advanced Research and Reviews 27, no. 1 (July 30, 2025): 974–88. https://doi.org/10.30574/wjarr.2025.27.1.2602.
[13] Cynthia Swann, “Unequal Ground: Rural America’s Legacy of Environmental Racism and Exploitation,” American Bar Association, November 5, 2025, https://www.americanbar.org/groups/crsj/resources/human-rights/2025-october/unequal-ground/.
[14] Ren, Shaolei, and Adam Wierman. “Mitigating the Public Health Impacts of AI Data Centers.” Harvard Business Review, November 5, 2025. https://hbr.org/2025/11/mitigating-the-public-health-impacts-of-ai-data-centers.
[15] Ren, Shaolei, and Adam Wierman. “Mitigating the Public Health Impacn foron forhavepollutantsts of AI Data Centers.”
[16] Ren, Shaolei, and Adam Wierman. “Mitigating the Public Health Impacts of AI Data Centers.”
[17] Ren, Shaolei, and Adam Wierman. “Mitigating the Public Health Impacon for,havepollutants,its,ts of AI Data Centers.”
[18] “Public Health Opportunities to Address the Health Effects of Air Pollution,” American Public Health Association, November 7, 2017, https://www.apha.org/policy-and-advocacy/public-health-policy-briefs/policy-database/2018/01/18/public-health-opportunities-to-address-the-health-effects-of-air-pollution.
[19] “AI Has an Environmental Problem. Here’s What the World Can Do about That.” UN Environment Programme
[20] Pereira, José Renato Laranjeira de. “The EU AI Act and Environmental Protection: The Case for a Missed Opportunity: Heinrich Böll Stiftung: Brussels Office – European Union.” Heinrich Böll Stiftung | Brussels office – European Union, April 8, 2024. https://eu.boell.org/en/2024/04/08/eu-ai-act-missed-opportunity.