By Sarah Hiller
Correspondent
You have probably used artificial intelligence today without realizing it. Not Googling a question and reading the Gemini response, or even opening up ChatGPT, within the regular apps you open every morning. Spotify deciding what song plays next. Google Maps rerouting you around traffic. Instagram choosing which post you see first. None of that feels blatantly like artificial intelligence, but it is, and it is constantly running. Data centers around the world are burning through electricity and water at an unsustainable rate.
Although, it is important to recognize that the conversation about AI‘s environmental impact has started. There are published findings and news outlets running the numbers and raising alarms — but awareness has not been translated into action. AI continues to expand into industries, and more corners of daily life. There is no serious policy movement, and no slowdown in sight.
According to a December 2025 study in Cell Reports Sustainability, AI data centers could produce between 32.6 million and 79.7 million tons of carbon dioxide in 2025, which is comparable to the annual emissions of a small country.
That same study found that the water required to cool those data centers this year could rival the entire global consumption of bottled water (312.5–764.6 billion L). For such an intangible product, it has a massive physical impact on our world.
A November 2025 study out of Cornell University, published in Nature Sustainability, projected that by 2030, AI could emit 24 to 44 million metric tons of carbon per year. Researchers believe this is comparable to 10 million cars on U.S. roads. Most of the electricity powering these data centers still comes from fossil fuels, not solar panels.
Some believe that the environmental costs of AI are somewhat overstated, and will correct themselves as technology improves and becomes more efficient. In an April 2025 report, the International Energy Agency argued that AI could ultimately be a net positive for the climate, writing that “the widespread adoption of existing AI applications could lead to emissions reductions that are far larger than emissions from data centres.”
The IEA’s point is that AI can be used to optimize energy grids, reduce industrial waste and accelerate clean energy development — potentially offsetting its own footprint.
That may be true in an ideal scenario. But it requires the benefits of AI to materialize at the same pace as its energy demands — and right now, there is little evidence that is happening. A May 2025 investigation by the MIT Technology Review found that AI is being integrated into “every corner of our digital lives,” and that the pace of adoption is far outrunning any efficiency improvements.
Sasha Luccioni, an AI and climate researcher at Hugging Face, told the publication that “generative AI tools are getting practically shoved down our throats and it’s getting harder and harder to opt out, or to make informed choices when it comes to energy and climate.” The optimistic version of the IEA’s argument depends on coordination that simply is not reflected in how the industry is currently operating.
As Luccioni said, stated plainly, what makes this harder to confront is that most of it is invisible. When AI is built into an app that you already use, there’s no moment when you choose to consume energy.
It’s not that you’re necessarily opting in, the AI is just there, running in the background as a feature, and the environmental cost is bundled in. Tech companies are not required to publicly report how much energy their AI system specifically uses.
The U.N. Environment Programme has called on governments to change that, recommending mandatory disclosure of AI’s environmental impacts. So far, those recommendations have gone largely unheeded.
This matters to students at the College, and students across the country, not just as individuals, but as future workers and consumers. Healthcare, finance, marketing, logistics and media are all integrating AI tools faster than any environmental standards can even be developed to control them.
It’s important to remember that technology itself is not the real enemy. AI has the tools to help address climate change, like modeling energy systems or detecting deforestation. It also has incredible potential in certain fields of work that may genuinely benefit its employees.
The issue is AI is expanding into every corner of daily life without transparency. Paired with a lack of regulation and serious scrutiny of the environmental tradeoffs, it is incompatible with the commitments governments and companies have made.
The conversation is already happening. The research already exists. What is missing is any sense of urgency from the people with the power to act on it. Companies should be required to report the energy and carbon emissions of their AI usage, data centers should meet renewable energy standards before expanding, and policymakers should stop treating AI’s environmental footprint as a distant future problem when it’s already happened.
AI will not slow down, it will continue to be developed and adopted. The question is whether the people producing it will take the environmental cost seriously.
