What Does AI Have to Do with the Environment, and How Can I Help?

If you watched the Superbowl this weekend, then you know AI – artificial intelligence –  has gone mainstream. Commercials from companies like Anthropic and Google’s Gemini AI aired alongside ads from big traditional brands like Budweiser. 

Ever since ChatGPT debuted in 2022, use of AI tools at work, school, and for entertainment has skyrocketed. Those tools need major data processing power to work – which means that as AI use grows, so does the environmental impact.

Compared with the rapid growth and vast size of the AI industry, individual actions may feel meaningless. But it turns out that your engagement and local action can make a big difference.

AI's Unique Environmental Impact

Every computer needs electricity to run. The Internet’s popularity has been growing since the 1990s. Back in 2006, leading search engine Google built its first standalone data center, a complex of buildings housing Internet servers to help accommodate its website traffic.

In that context, AI may seem like just the next step in a progression. But AI is exponentially more resource-intensive than its predecessors. As this helpful illustrated article from the New York Times explains, the chips that power AI are called GPUs, and they need at least twice as much wattage as the CPUs that have traditionally been used to power web servers and computers. 

Speed is another factor. When you have a back-and-forth “conversation” with ChatGPT, each query prompts the tool to process vast amounts of information in fractions of a second, which means many GPUs need to work together to keep up the pace. 

To meet this demand, companies reliant on AI are building data centers – huge buildings the size of multiple football fields that store thousands of GPU computers. These data centers impact the environment in a number of ways, including:

• Energy usage - The data centers the company OpenAI alone plans to build will use as much energy as the 3 million households in Massachusetts, according to the New York Times article linked above. By 2028, data centers could account for 12% of all U.S. energy consumption, according to the Lawrence Berkeley Lab.
  

• Water usage - To ensure the GPU computers don’t overheat and melt down, cooling units circulate chilled water. National Public Radio reports that a mid-sized data center uses roughly 300,000 gallons per day, equivalent to 1,000 households. Additionally, these cooling units run 24/7, causing noise problems if the data center is close to residential neighborhoods.
  

• Pollution - With data centers putting pressure on local energy grids, utilities are turning to gas- and coal-fired power generation to keep up with demand. In addition, backup diesel generators – which can also be used as supplemental power when demand is high – release nitrous oxide and other toxins, according to the Harvard Business Review.

What You Can Do

The companies that own data centers can seem too big for personal activism to make a difference. But there are steps you can take. 

• While AI companies are global, the process of building new data centers is largely local. They must adhere to local permitting regulations, submit plans to local boards, and follow environmental reporting requirements. Follow local government, know what projects are being proposed near you, and collaborate with community environmental organizations to voice concerns.
  

• Petition local, state, and federal representatives to hold companies accountable. Environmental impact metrics and reporting should be standardized, and transparency required. Fines for violations should be significant enough for companies to take notice. In addition, companies should have to pay their own way when it comes to energy use; petition your elected representatives to push back if AI power costs end up affecting your utility bill.
  

• Support scientific research and evidence-backed development of clean energy to power tomorrow’s technology. While the AI hype may ebb, the fact is that electronics are here to stay and they require power to run. If you have the means, donate to scientific organizations that prioritize clean energy, and vote for elected official who prioritize science.
  

• Push back on the “AI is inevitable” storyline. It’s true that Meta won’t notice your personal decision to leave Facebook, but you can do your part to deflate AI hype and resist the argument that we all need to start using AI to help it improve. Encourage friends and family to spend more time face to face. Encourage discussion about the impact of AI on artists and writers whose works are being used as “training data,” often without compensation. At the office, track whether AI tools truly save time; consultancy McKinsey recently found that 80% of companies using AI report no significant benefits. 

AI’s impacts on the environment are significant, and are poised to grow. But by organizing your community to push back, you can help create collective momentum to change the trajectory of future technology.

Sources

Brookings, "AI, Data Centers, and Water," 2025 - https://www.brookings.edu/articles/ai-data-centers-and-water/

Data Center Watch - organization tracking opposition to data centers - https://www.datacenterwatch.org/

New York Times, "How AI Is Changing the Way the World Builds Computers," 2025 - https://www.nytimes.com/interactive/2025/03/16/technology/ai-data-centers.html

NPR, "Data centers, backbone of the digital economy, face water scarcity and climate risk," 2022 - https://www.npr.org/2022/08/30/1119938708/data-centers-backbone-of-the-digital-economy-face-water-scarcity-and-climate-ris

The Guardian, "Rage Against the Machine: A California Community Rallied Against a Giant Datacenter - and Won," 2026 - https://www.theguardian.com/us-news/2026/feb/07/california-monterey-park-stop-datacenter-construction