Artificial Intelligence (AI) is on the rise. It’s transforming the world we live in—from entertainment to self-driving cars. AI promises efficiency and innovation, and as students, we increasingly rely on AI tools. We use it for proofreading, brainstorming, and helping us pass our classes. However, it is important to be informed about the tools we utilize. What are the hidden costs?
A big concern about AI models such as ChatGPT is their energy consumption. They rely on big data centers, and in 2022, it was found that data centers for AI and cryptocurrencies (which is another dilemma we will not go into here) accounted for around 3% of the U.S. energy demand (Goldman Sachs, 2024). This energy demand is also expected to increase in the coming years. It has been estimated that AI’s energy demand could increase to around 8% of the U.S.’s total energy consumption by 2030 due to increased data center power consumption (Goldman Sachs, 2024). The training of these models is also extremely energy intensive.
“The training process for a single AI model (…), can consume thousands of megawatt-hours of electricity and emit hundreds of tons of carbon. This is roughly equivalent to the annual carbon emissions of hundreds of households in America.”
—Ren, S., & Wierman, A. (2024).
And the energy consumption doesn’t stop when the training is done. It is estimated that an inquiry to ChatGPT uses five times more energy than a Google search (MIT News, 2025a).
There are also other concerns beyond energy consumption. Data centers require a lot of water for cooling, and soon, AI-related data centers will consume around six times more water than the country of Denmark, which has a population of 6 million people (MIT News, 2025a; UNEP, n.d.).
Is the world doing anything? While the impact of AI is significant, there are steps being taken to mitigate its footprint (MIT News, 2025b). Researchers are developing more energy-efficient models, and tech companies are starting to invest in renewable energy to power data centers. There is also a focus on training the models when it’s colder to reduce the need for water cooling in data centers (MIT News, 2025b). There is still a long way to go, and therefore, it is important for the issue to stay in focus.
As students, we can think a little bit more about how we utilize AI—do we really need AI for this specific task? Maybe we can switch out the ChatGPT inquiry with a Google search? AI has the potential to help solve some of today’s biggest challenges—from optimizing renewable energy to climate modeling. If managed sustainably, AI could be part of the solution instead of the problem.
References:
Goldman Sachs. (2024). AI poised to drive 160% increase in power demand. Goldman Sachs. https://www.goldmansachs.com/insights/articles/AI-poised-to-drive-160-increase-in-power-demand
MIT News. (2025). Explained: Generative AI’s environmental impact. Massachusetts Institute of Technology. https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117
MIT News. (2025). Q&A: The climate impact of generative AI. Massachusetts Institute of Technology. https://news.mit.edu/2025/qa-vijay-gadepally-climate-impact-generative-ai-0113
Ren, S., & Wierman, A. (2024). The uneven distribution of AI’s environmental impacts. Harvard Business Review, 15. https://hbr.org/2024/07/the-uneven-distribution-of-ais-environmental-impacts
UNEP. (n.d.). AI has an environmental problem—Here’s what the world can do about it.United Nations Environment Programme. https://www.unep.org/news-and-stories/story/ai-has-environmental-problem-heres-what-world-can-do-about
