AI and the Environment: Considerations and Concerns

AI and the Environment: Considerations and Concerns

Developments in Generative AI technology and the promise of further enhancements and innovations gesture toward a near future filled with computational possibilities and solutions to various problems in multiple sectors. However, at what environmental cost do these advancements come? With the rise of popular AI programs like Gemini, ChatGPT, etc, and the release of these models to the general public, there is a growing awareness of the environmental costs of this new technology. While AI can arguably be used to help us find solutions to environmental problems, there are very real costs that accompany the creation and application of this technology. Some of these environmental impacts to consider include infrastructure needs, energy usage, E-waste and resource extraction, carbon emissions, and impact on local communities.

image-20250418-170658.png

Infrastructure Needs: E-Waste and Extraction

The hardware and software that supports AI during all stages of its lifecycle can be referred to as AI Infrastructure. As AI technologies become more advanced, the need for sophisticated infrastructure also increases. Hardware and other components that support massive data storage and management, computational abilities, and data processing frameworks are critical in ensuring the smooth operation of AI. Production of these special types of hardware (like GPUs, TPUs, and specialized chips) will ultimately lead to a higher volume of E-Waste. In addition, producing these items will require continued resource extraction of rare earth minerals (for example, lithium and cobalt). 

The environmental costs to the local environments where these minerals are mined can vary in their degree of impact but can include deforestation, pollution of local resources, and habitat destruction. The communities that are the most affected by mineral extraction also tend to be communities that benefit the least from the application of AI, bringing forth a larger question of social inequity and access to AI. 

Energy Consumption: Electricity and Water

Training large AI models requires an incredible amount of computational power, which in turn requires a large amount of energy. Data centers consume high amounts of electricity, which can strain already overburdened electric grids. For example, training GPT-3 reportedly required 1,287 MWh of electricity and emitted roughly 552 tons of CO2. The International Energy Agency estimates that the “global electricity demand from data centers could double between 2022 and 2026, fueled in part by AI adoption”. 

Additionally, vast amounts of water are required to cool data centers and prevent malfunctions from overheating. For example, Google’s AI-related data centers used billions of liters of water in 2022 alone. If a data center is located in an area where this resource is already scarce (the American Southwest, for example), this could put a strain on local water budgets and impact surrounding communities. 

Carbon Footprint

AI-driven cloud computing services generate significant C02 emissions. Data centers and cloud computing contribute to nearly 3% of global energy-related greenhouse gas emissions. As AI continues to develop and grow in popularity, these numbers can be expected to rise.

On the one hand, AI could potentially offer solutions to environmental problems, through the application of its technology on precision farming and responsible land/water/crop usage, smart urban planning, and creating efficient pathways for the expansion of public transportation (just to list a few possibilities); on the other hand, these potential innovations will not come in to being simply by being devised by AI: like any solutions proposed by AI, organizations and companies–that is, people–are responsible for actually implementation, which can be strengthened or stymied by social, political, and economic factors. 

Further Reading:

United Nations Environment Programme. "AI Has an Environmental Problem. Here's What the World Can Do About That." UNEP, 21 Sept. 2024,https://www.unep.org/news-and-stories/story/ai-has-environmental-problem-heres-what-world-can-do-about .

Ren, Shaolei and Weirman, Adam. "The Uneven Distribution of AI's Environmental Impacts." Harvard Business Review, 15 July 2024, 

https://hbr.org/2024/07/the-uneven-distribution-of-ais-environmental-impacts

Zewe, Adam. "Explained: Generative AI's Environmental Impact." MIT News, 17 Jan. 2025,https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117 .