OIT pledges to remain environmentally responsible as AI use increases 

AI use has expanded drastically in higher education and has led to the assessment and exploration of AI’s impact on teaching, learning and researching.  

The Office of Information and Technology has pledged to uphold sustainability standards. According to Greg Williams, OIT deputy chief information officer, OIT offsets their power consumption from one computational resource to another. 

“While OIT consumes a lot of power, we are very environmentally responsible … for example, we have cut the number of computers in the library from about a half, allowing us to use virtual machines, which are far more efficient at about a tenth of the energy costs,” Williams said. 

These virtual machines allow students, faculty and staff to access university desktop hardware from their laptops, minimizing the need of campus data servers and increasing efficacy. 

UCCS staff and faculty have implemented classroom policies to encourage responsible and ethical AI use. The OIT AI page outlines AI tools like Microsoft Copilot and Adobe Firefly and explains the guidelines for using AI safely.  

According to the page, “CU is in the process of establishing guidelines and compliance requirements for using AI — with the complexity of this technology, it may take some time. And without proper security controls, AI technology can become susceptible to privacy, confidentiality and security threats.” 

Carbon emissions, water, energy consumption and electronic waste have increased due to AI use. Carbon emissions from non-renewable electricity and the consumption of millions of gallons of fresh water causes environmental sustainability concerns due to the global climate crisis, according to a study from the University of Illinois.  

The use of generative AI, such as ChatGPT, has risen since its launch in 2022. A 2024 study from the University of California Irvine found that over half of US college students use ChatGPT for general purposes and writing at least monthly.  

Data centers have had to increase energy and water usage to meet the demands of AI research. AI, Algorithmic and Automation Incidents and Controversies estimated that ChatGPT uses about 500 milliliters of water for every 5–50 prompts. 

Dylan Harris, assistant professor in the department of geography and environmental studies, explained that AI data centers require cooling technology, including air-conditioning systems that reduce overheating with water or refrigerant and evaporative cooling using evaporated water. 

AI data centers’ rapid development in regions with cost-efficient water sources, such as waste-water and recycled water, strains the local water supply in drought prone regions, such as Colorado, according to Harris.  

“Data centers around the US are either located in energy stress regions or water stress regions.  They’re located basically in the worst places possible, and that’s partially due to those locations having either cheap energy or cheap water,” Harris said. 

According to NBC News, global data center infrastructure increased by 6% from 2020 to 2021, followed by an estimated 3-4% annually within the last three years. This growth correlated with the drought conditions in the west in 2020. 

Students can learn more about AI data practices on OIT’s AI page.