William Shakespeare coined the phrase to fight “fire with fire” in 1623’s The Life and Death of King John.[1] A mere 402 years later, the same phrase used to describe English succession wars may apply to a newly formed, though equally existential, question of how to deal with climate change using a technology with a well-known environmental impact, Artificial Intelligence (AI).[2] The environmental impact of AI has been well documented: from the rare-earth metals needed to develop powerful AI-capable microchips, to the energy needed to run AI computer banks, to the water needed to keep those computers cool and operational, AI tools’ environmental costs are significant.[3] However, despite these costs, AI may yet provide an invaluable tool in fighting climate change and pollution to protect our planet.
In the last decade, AI has become a seemingly ubiquitous answer to any question. The Trump Administration recently issued an executive order[4] and Request for Information[5] asking for commentary on how to safeguard the United States’ position as a leader in AI. Given this push from both industry and government actors to advance the use of AI across sectors, there is an invaluable opportunity to center AI tools in the fight against climate change.
Minimizing Impacts
As already noted, certain environmental costs appear to be baked-in to the creation of AI models and development of AI tools. However, these costs can be minimized. While significant research is necessary to replace the rare-earth metals currently necessary to build microchips, AI tools are actively being used by researchers to design more efficient and powerful chips.[6] The future of this research could mean that systems can be run using fewer chips, producing less e-waste, and those chips can be made more efficiently.
Other parts of the AI ecosystem can, and are, being made greener already. For instance, based on public reporting, companies leading in the development of AI, including Google and Microsoft have announced investments in nuclear energy – including small-scale reactors – to provide “a particularly reliable form of carbon-free energy” which is “twice as reliable [meaning consistent] as coal and natural gas,” as well as other clean energy forms like wind and solar.[7] Similarly, while traditional data center cooling can use “anywhere from one million to five million gallons of water per day[,] as much as a town with 10,000 to 50,000 residents,” due to the intense heat of servers evaporating the water used to cool them, Microsoft has recently introduced advanced “zero-water evaporated designs” for data center cooling needs.[8] This technology functions by using a “closed loop” and recapturing water as it evaporates. It, along with using reclaimed and recycled water, can be used to significantly decrease the impact of running the servers necessary to create and maintain AI models. Although not a complete answer, these developments show significant progress in making AI a smaller net cost in environmental terms.
Using AI to Fight for the Environment
Beyond helping make the AI ecosystem itself more efficient, AI presents a valuable opportunity to track and fight climate change, increase sustainability efforts across varying sectors, and prepare for and respond to disasters more effectively.
The World Economic Forum recently released a report describing ways in which AI is helping track climate change’s effects, including: analyzing ice melt in the Arctic, tracking deforestation across forest systems (and helping reforest the Amazon), making recycling more effective and cost efficient, and clearing waste from the oceans, among others.[9] According to the World Economic Forum report, waste is responsible for 16% of global greenhouse gas emissions through production of methane and other greenhouse gases. A software startup in London “has developed an AI system that analyzes waste processing and recycling facilities to help them recover and recycle more waste material[,] tracked 32 billion waste items across 67 waste categories in 2022, and . . . identifie[d] 86 tonnes [roughly 95 U.S. tons] of material on average that could be recovered but is being sent to landfill.”[10] In another waste management example, a scientist with the National Institute of Standards and Technology has recently used “machine learning algorithms and other scientific techniques” to train a computer “to identify . . . plastic based on how similar [a] new [near-visible infrared light (NIR)] signal is to the NIR signals of other plastics” which helps “identify the material in a soda bottle, know that it’s different from the makeup of a takeout container, and separate them” automatically with less human intervention.[11] Tools like this can make recycling more efficient and less costly for localities, which will improve both the effectiveness of local recycling programs and allow municipalities to recycle more for the same cost.
In the energy industry, AI technology is being used to improve efficiency in the production and transmission of electricity. Similar to tracking drought conditions or ice melts, AI can use satellite imagery and sensor data to analyze weather patterns, and compare them with electricity demand patterns.[12] This can help energy producers optimize the operation of renewable energy sources like wind and solar power, increasing efficiency and aiding in grid integration.[13] It also improves the bottom line of the companies utilizing it – with the possibility of reducing “operational costs by up to 15%, and boost productivity by 10%” for energy producers.[14] The Abu Dhabi state-owned oil company ADNOC, for instance, used AI energy-saving efforts which “generated $500 million in value and reduced carbon emissions by about a million tonnes [1,102,311 U.S. tons] – the equivalent of removing around 200,000 gasoline-powered cars from the road.”[15] These types of efficiency improvements – even without switching to green power – show that AI can provide impressive tools to reduce emissions.
Electricity delivery is also primed for AI-driven efficiency improvements. The power grid is “often described as the most complex machine ever built” because of its vastness.[16] The Argonne National Laboratory partnered with a grid operators in 2019 to use machine learning and AI models to help with the “complex mathematical calculations that predict how much electricity will be needed the next day and try to come up with the most cost-effective way to dispatch that energy.”[17] The AI systems could make those calculations 12 times faster than without AI, reducing the time it took from almost 10 minutes to 60 seconds, and given that “system operators do these calculations multiple times a day, the time savings could be significant.”[18] Researchers at the University of Virginia have also begun using new AI models using “graph neural networks” to “improve power flow analysis – the process of ensuring electricity is distributed safely and efficiently across the grid.”[19] These tools can help the grid adapt to various configurations and changes – like power line failures.[20] Moreover, they can help determine “how much power should be generated from different sources,” which is increasingly important as “renewable energy sources introduce uncertainty in power generation and distributed generation systems” and electrification of transportation systems increases demand on the grid.[21]
AI can also shows potential to help decarbonize agriculture and increase its efficiency and resiliency. Globally, agriculture was responsible for 22% of 2019 global greenhouse gas emissions according to the EPA.[22] A recent McKinsey report identified agriculture as “particularly well suited” for AI deployment because of “high volumes of unstructured data, significant reliance on labor, complex supply chain logistics,” and other factors.[23] The National FFA Organization also released a report discussing how AI can help farmers improve productivity on their acreage by advising farms on “what to plant, when to plant it, when to fertilize, [and] when to harvest,” among other recommendations.[24] It can also advise on impending crop stresses from uncontrollable conditions like weather, water stress, pests, or disease.[25] Most importantly for the climate fight, AI can “help farmers adopt regenerative agriculture practices, including reduced till, cover cropping[,] and nutrient management, by helping them analyze the impact on soil carbon and financial profitability[, and] . . . can indicate where fertilizer is needed (which conserves resources and reduces nitrous oxide emissions) and reduce water use through precision irrigation.”[26] During adverse weather conditions, like drought, AI can also play a key role. A recent United Nations report emphasized that AI-powered applications have helped Kenyan pastoralists “brace for drought. With data from global meteorological stations and satellites sent to their mobile phones, herders can plan ahead, better manage their livestock[,] and save hours of scouting for green pastures.”[27] These applications increase efficiency, decrease over-farming, improve regenerative agricultural practices, and protect against the negative impacts of a less predictable climate.
Finally, AI can help prepare for and protect against disasters. In a previous blog post, we have discussed the coming climate refugee crisis. The United Nations has cited AI’s potential to help with a challenging future of climate disasters. UN agencies in Burundi, Chad, and Sudan have used AI tools to “investigate past environmental change around displacement hotspots and deliver future projections to inform adaptation measures and anticipatory action for integration in humanitarian programming.”[28] Moreover, AI-enhanced modelling and predictive technologies “can help communities and authorities to draft effective adaptation and mitigation strategies” even absent acute disasters.[29]
Domestically, utilities have been implementing AI tools for risk mitigation as well. Overgrown trees are “a leading cause of blackouts” because branches can fall on electric lines or may spark fires.[30] PG&E has used AI and machine learning along with drones to accelerate inspections, “identify areas requiring tree trimming,” or pinpoint “equipment that needs repairs,” and prioritize work.[31] Thus, AI is actively helping to fight fire with fire (prevention). Moreover, at least one startup has begun using AI to analyze utility companies’ historical data on energy equipment performance and “combine[] it with global climate models to predict the probability of grid failures resulting from extreme weather events, such as snowstorms or wildfires.”[32] In 2023, the EPA released a report identifying the means by which AI can improve post-disaster recovery – including improving communication and collaboration among stakeholders, improving the safety and efficiency of recovery and response activities, and helping to categorize the types of debris left over after disaster strikes.[33] These tools are increasingly necessary as climate events and disasters become more prevalent and severe.
In short, despite its environmental impacts at present, AI provides an important tool to fight climate change and mitigate its impacts. If used correctly, the environmental costs associated with creating AI models are likely to be outweighed by their positive impacts on environmental causes – helping to fight fire with fire.
*Nathaniel Schetter also contributed to the publication of this blog post.
[1] William Shakespeare, King John act 5, sc. 1, l. 49.
[2] This blog post adopts the National Institute of Standards and Technology definition of AI, “A machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments.” Artificial Intelligence, Nat’l Inst. of Standards & Tech., https://csrc.nist.gov/glossary/term/artificial_intelligence (last visited Apr. 18, 2025).
[3] AI Has an Environmental Problem. Here’s What the World Can Do About That., UN Env’t Programme (Sept. 21, 2024), https://www.unep.org/news-and-stories/story/ai-has-environmental-problem-heres-what-world-can-do-about.
[4] Exec. Order No. 14179, § 1, 90 Fed. Reg. 8741, 8741 (Jan. 23, 2025) (“With the right Government policies, we can solidify our position as the global leader in AI and secure a brighter future for all Americans.”).
[5] Request for Information on the Development of an Artificial Intelligence (AI) Action Plan, 90 Fed. Reg. 9088 (Feb. 6, 2025).
[6] Caroline Delbert, AI Designed Computer Chips That the Human Mind Can’t Understand., Popular Mechs. (Jan. 31, 2025), https://www.popularmechanics.com/science/a63606123/ai-designed-computer-chips/.
[7] Julia Shapero & Rachel Frazin, Big Tech Turns to Nuclear Energy to Power AI Boom, The Hill (Oct. 16, 2024), https://thehill.com/policy/technology/4937529-google-amazon-nuclear-power-energy-ai/.
[8] Nicole Cappella, Microsoft Unveils Closed Loop Water Cooling for Data Centres, Techerati (Dec. 12, 2024), https://www.techerati.com/news-hub/microsoft-unveils-closed-loop-water-cooling-for-data-centres/.
[9] Victoria Masterson, 9 Ways AI Is Helping Tackle Climate Change, World Econ. F. (Feb. 12, 2024), https://www.weforum.org/stories/2024/02/ai-combat-climate-change/.
[10] Id.
[11] Bradley Sutliff, From Trash to Cash: How AI and Machine Learning Can Help Make Recycling Less Expensive for Local Governments, Nat’l Inst. of Standards & Tech. (May 15, 2024), https://www.nist.gov/blogs/taking-measure/trash-cash-how-ai-and-machine-learning-can-help-make-recycling-less-expensive.
[12] Kara Anderson, How Can Artificial Intelligence Help Tackle Climate Change?, Greenly: Leaf (Feb. 12, 2024), https://greenly.earth/en-us/blog/industries/how-can-artificial-intelligence-help-tackle-climate-change.
[13] Id.
[14] Ibrahim Al-Zu’bi, Energy and AI: The Power Couple That Could Usher in a Net-Zero World, World Econ. F. (Jan. 29, 2025), https://www.weforum.org/stories/2025/01/energy-ai-net-zero/.
[15] Id.
[16] June Kim, Four Ways AI Is Making the Power Grid Faster and More Resilient, MIT Tech. Rev. (Nov. 22, 2023), https://www.technologyreview.com/2023/11/22/1083792/ai-power-grid-improvement/.
[17] Id.
[18] Id.
[19] New AI Model Could Make Power Grids More Reliable Amid Rising Renewable Energy Use, Univ. of Va. Sch. of Eng’g & Applied Sci. (Oct. 24, 2024), https://engineering.virginia.edu/news-events/news/new-ai-model-could-make-power-grids-more-reliable-amid-rising-renewable-energy-use.
[20] Id.
[21] Id.
[22] Global Greenhouse Gas Overview, U.S. EPA (Mar. 31, 2025), https://www.epa.gov/ghgemissions/global-greenhouse-gas-overview.
[23] Daniela Nuscheler et al., From Bytes to Bushels: How Gen AI Can Shape the Future of Agriculture, McKinsey & Co. (June 10, 2024), https://www.mckinsey.com/industries/agriculture/our-insights/from-bytes-to-bushels-how-gen-ai-can-shape-the-future-of-agriculture.
[24] How AI Can Impact Agriculture, FFA (Aug. 18, 2023), https://www.ffa.org/technology/how-ai-can-impact-agriculture/.
[25] Id.
[26] Id.
[27] Explainer: How AI Helps Combat Climate Change, UN News (Nov. 3, 2023), https://news.un.org/en/story/2023/11/1143187.
[28] Id.
[29] Id.
[30] Kim, supra note 16.
[31] Id.
[32] Id.
[33] Timothy Boe, U.S. EPA, EPA/600/R-23/120, The Current State of Artificial Intelligence in Disaster Recovery: Challenges, Opportunities, and Future Directions (2023).
[View source.]
