{shortcode-213ef4edf34cd90bbe7f6492a36757114418d89b}

As Harvard students enjoy University-provided access to generative artificial intelligence, communities around the country are suffering.

The rapid expansion of infrastructure powering artificial intelligence has wrought serious environmental damage Harvard must align its guidance for AI usage with the technology’s devastating impact on the environment, discouraging its use for non-essential, non-academic purposes.

In our hometown of Memphis, Tennessee, the impact of AI isn’t just academic. The world's largest supercomputer (per its creators), aptly named “Colossus,” currently stands on hundreds of acres near the predominantly Black community of Boxtown. Its company is contracted to burn through enough electricity to power 100,000 homes a year, consume up to a million gallons of water per day, and is potentially leaking toxic chemicals into our air.

The environmental impact is both a looming threat and an immediate concern, leaving Memphians scrambling for a solution.

And this isn’t just a Memphis problem. Data processing plants are popping up across the country, fueled by massive demand for AI.

There are several consequences to this trend. A recent UC Riverside and Caltech study estimates that the total public health impact of data centers in the U.S. in 2030 is up to $20 billion per year, comparable to the impact of on-road emissions in California. Concerningly, some of the data centers using large amounts of water for cooling are operating in regions already stressed for water.The MIT technology review concluded that in 2024, American data centers used around 200-terawatt hours of electricity — enough to power the country of Thailand for a year.

Yet Harvard’s Guidelines for Using ChatGPT and other Generative AI Tools, along with many of Harvard’s AI policies and resources, focus primarily on concerns like “information security and data privacy, compliance, copyright, and academic integrity” while omitting environmental implications.

Meanwhile, students and faculty use AI casually and frequently, often for non-essential creations: proctor door tags, replying to simple emails, or summarizing facts that could easily be searched. These seemingly-harmless uses are contributing to the expansion of AI supercomputers. By taking a public stance on the clear environmental impacts of this technology, Harvard can support those most affected, many of whom are Harvard students, and create a more conscientious group of future AI users.

We’re not calling for an outright ban on AI. These tools can present a unique opportunity to further academic inquiry and experimentation. But, as students and academics, we have a responsibility to use them thoughtfully and only when they add value to academic inquiry.

We should foster a campus culture that recognizes and challenges the hidden cost of convenience. Harvard must revise its AI guidelines to include environmental considerations, encouraging greater awareness in the classroom.

As students, we must reflect on our AI habits. As you use AI for tasks that can be done manually, consider the climate cost of your actions. Use the vast array of academic resources available to you before you turn to AI. Harvard provides for Course Assistants, Teaching Fellows, the writing center, and House tutors that support students, encourage critical thinking, and are much more environmentally conscious. AI is simply not a necessity for a school with so many resources.

As Memphians, we carry our city’s struggles with us anywhere we go, including here on campus. It is frustrating to see our peers rely on AI without regard for the impact it has on neighborhoods close to those we grew up in. We are consistently angered by the injustices that our community experiences while we remain untouched here in Cambridge, removed from their struggle.

As long as Harvard hopes to be a sustainable institution, we must address these pressing concerns. We shouldn’t ask communities all over the nation to shoulder the cost of our convenience. Harvard must lead the way in using AI sparingly, thoughtfully, and only when it truly advances academic inquiry.

L.A. Karnes ’28, a Crimson Editorial Editor, lives in Mather House. Ira Sharma ’28, a Crimson Editorial editor, lives in Mather House.