According to The 2025 AI Index Report, the use of Artificial Intelligence (AI) has increased globally to 21.3% across 75 countries since 2023, representing a ninefold increase since 2016.
Companies heavily rely on AI; around 78% of businesses are using it for at least one business function (McKinsey & Company). The amount of AI usage can impact the environment negatively through high energy consumption from data centers, which leads to a large amount of greenhouse gas emissions, resulting in climate change.
High energy consumption from data centers contributes to climate change because of its reliance on electricity, often acquired from fossil fuels. When fossil fuels are burned, they release nitrogen oxides into our atmosphere. Excess nitrogen in the atmosphere contributes to pollution, climate change and oxygen-deprived aquatic zones.
Using AI requires massive data centers and powerful hardware. The rapid expansion of AI has led to a boom in data centers, placing a strain on local communities and habitats for animals.
While high energy consumption is an issue, some companies, like Google, that involve AI are switching to a net-zero carbon goal.
“Google’s energy demand for a Gemini prompt itself equates to about five drops of water. Another way you can conceptualize that is about 30 prompts equate to a bottle of water,” Andrew Fenstermaker (he/him), Instructional Technology Coordinator for ICCSD, said.
The Duane Arnold Energy Center in Palo, Iowa, was closed in 2020, but could be running by early 2029 after being purchased by Google to be run as a “24/7 carbon-free energy source” to help power Google’s growing AI infrastructure in Iowa. This is meant to help meet the energy demands of data centers and artificial intelligence. Although according to sources, Google’s alleged goal to be net-zero in Carbon emissions by 2030, as posted on the front page of their website, has been quietly removed.
“Services like Cloud from Amazon are already using big data centers, which use a fair amount of energy, and then they use water to cool down all the computers, and that’s an environmental impact,” Juan Pablo (he/him), professor at the University of Iowa in the Department of Computer Science, said.
Data centers typically cool down their computers by using local freshwater supplies. A lot of new data centers are being built in places with water scarcity, which increases competition for resources for local agriculture and the community. The water scarcity and competition for resources can threaten local ecosystems.
Big companies that have AI models like Llama, ChatGPT, Gemini and many others have the capital resources to invest in massive data centers. They build these data centers because they are needed to create the models they use. The U.S. accounted for approximately 176 tWh (Terawatt-hours) in 2023, which is 4.4% of annual energy use, and is expected to increase to 12% of the nation’s energy by 2028.
“When [companies] release new models, they become more and more efficient. So the demand for energy and the computing power needed continues to be more efficient,” Fenstermaker said.
Building models uses a lot of energy and resources, but even more resources are stripped from the simple questions that people ask AI.
“One part of it is building the models, but it turns out that most of the energy is used in all the queries that we ask the generative AI systems to make for us. Individually, they aren’t so bad for small questions or images; the tricky part is the computer applications that constantly run queries. For example, you call a company and when they answer, it’s an AI system speaking to you,” Pablo said.
Companies use AI to assist with certain business needs. For example, AI allows for more accurate predictions of storms and other weather events, helping communities prepare promptly. But when AI is constantly running for a business, the data centers need to be cooled, and this consumes large amounts of water.
AI is rapidly growing due to the exponential increase in computing power and breakthroughs in machine learning algorithms. This rapid growth is predicted to lead to mass job displacement, social inequality, security risks and is already having a negative environmental impact.
AI systems can discriminate via unlawful and implicit biases on a system-wide scale. It can also provide false answers to queries.Experts stress that it is important to think for yourself instead of turning to AI, as it has negative impacts on people’s intellect and the environment.
