WASHINGTON, Nov 7 (Reuters) – (This report from Nov. 7 has been corrected to state that the date in paragraph 17 is 2017, not “last year.”)
Environmental organizations around the world are using artificial intelligence to analyze climate data, track forest fires, and count endangered wildlife—embracing the technology while simultaneously warning that it is driving a surge in electricity demand, water consumption, and greenhouse gas emissions.
This contradiction compels these organizations to justify their use of artificial intelligence in carrying out their mandates to address climate change or protect natural resources.
“We recognize that artificial intelligence can carry environmental costs, so we need to be thoughtful about how we use it,” says Ali Swanson, lead for artificial intelligence at the U.S.-based non-profit organization Conservation International. “But at a time when the climate and biodiversity crises are becoming increasingly urgent, AI can help us act faster and make better use of the information we collect.”
The organization has recently begun using AI to assess hotspots detected by satellites and to send real-time alerts about potential wildfires to community forest patrols in remote, environmentally sensitive areas of Cambodia.
Reuters interviewed 15 major environmental, conservation, and development organizations about their use of artificial intelligence. All of them reported using the technology in some capacity, while also expressing concern regarding the environmental impact of the data centers that power AI—facilities packed with servers that consume vast amounts of electricity and water for cooling.
The trade-off between the benefits of artificial intelligence and its significant risks poses a global challenge for individuals, governments, and industries alike. At this month’s UN Climate Change Conference (COP31), the host nation, Turkey, will call on governments, corporations, and other organizations to sign a pledge outlining principles to ensure that AI is developed and deployed in a sustainable manner—a context in which these trade-offs are likely to take center stage.
UN Secretary-General António Guterres has urged major AI companies to publicly disclose the environmental impact of their systems.
Researchers at the UN University Institute for Water, Environment and Health estimate that by 2030, the global footprint of data centers could more than double, reaching an area roughly twice the size of Jakarta, the sprawling Indonesian metropolis.
Meanwhile, the International Energy Agency (IEA) projects that by 2030, global data centers will consume approximately 945 terawatt-hours of electricity annually, resulting in carbon dioxide emissions equivalent to over 300 million tonnes. This figure represents double the current level of consumption—an amount roughly comparable to the total annual electricity usage of Japan.
The organization also forecasts that global water consumption by data centers will rise from approximately 560 billion liters (147.9 billion gallons) in 2023 to about 1.2 trillion liters (317 billion gallons) annually by 2030.
Tech giants including Microsoft (MSFT.O), Amazon (AMZN.O), and Alphabet’s (GOOGL.O) Google have reported rising emissions—partly driven by the expansion of data centers—raising questions about whether they can meet their long-term climate goals.
Amazon told Reuters that its artificial intelligence products are among the most sustainable available, citing the company’s procurement of low-carbon energy and its efforts to use electricity and water efficiently.
Neither Microsoft nor Google responded to requests for comment.
Breakthroughs once thought impossible
Many organizations interviewed by Reuters believe that the current benefits of artificial intelligence outweigh its costs, as the technology enables tasks that were previously difficult, prohibitively expensive, or even impossible to perform.
Save the Elephants, a partner of the Wildlife Conservation Network in Kenya, is leveraging machine learning to improve population surveys and study how elephants communicate.
Frank Pope, the organization’s CEO, noted that an aerial survey of Kenya’s largest protected area in 2017 yielded approximately 200,000 images, which took nine people nine months to review manually. Today, an artificial intelligence model—trained on the NVIDIA GeForce graphics platform—can complete a similar analysis in just a few hours.
“The results were eye-opening,” Pope said. “We discovered that human-led surveys had missed about a third of the elephants, roughly half of the zebras and giraffes, and three-quarters of the antelope population.”
He explained that more accurate wildlife data helps in securing funding for conservation efforts.
“Ultimately, these tools make nature measurable—and therefore more investable,” he said. “If we want to protect biodiversity on a large scale, this will be a crucial shift.”
At the institute founded by the late conservation icon Jane Goodall, researchers are using artificial intelligence to digitize decades of field notes on chimpanzee behavior from Gombe National Park in Tanzania.
The project was initially expected to take 13 to 15 years, as the records were handwritten in the local Kiha language and scientific shorthand.
Lillian Pintea, Vice President of Conservation Science at the Jane Goodall Institute, stated that thanks to a tool developed in collaboration with Amazon Web Services, these notes can now be scanned, translated, and organized in minutes—reducing the total time required to approximately two years.
Pintea noted that Goodall herself supported the use of AI to advance conservation science, provided that humans remain firmly in control of how the technology is applied.
“AI technology should be viewed as a human-centered tool,” Pintea said.
Limiting Environmental Impact
Some organizations are exploring various approaches to mitigate the environmental impact of using artificial intelligence. The World Resources Institute (WRI), a research organization that leverages artificial intelligence to accelerate climate and environmental analysis, states that it limits resource-intensive AI applications—such as the generation of videos and images—and caches AI query results so that identical searches do not require redundant processing.
The Nature Conservancy, based in the United States, has adopted a similar approach. The organization uses AI to analyze video footage captured on commercial fishing vessels to track the catch of various fish species. Rather than uploading vast amounts of video to a centralized data center, they employ AI-powered digital cameras installed directly on the boats to process the information, thereby reducing both energy consumption and data transmission requirements, explains Vienna Saccomanno, a senior scientist at the organization.
“Implementing edge computing requires a significant amount of engineering,” she says. “But it is well worth it to reduce our carbon footprint and gain these sustainability-related insights.”
Greenpeace International notes that it also utilizes certain AI tools, primarily because the software platforms it relies upon incorporate AI features.
The organization states that it is developing internal guidelines emphasizing transparency, human oversight, and the “conscious” use of AI. However, given the environmental impact of AI, it generally prefers to avoid the technology and strongly opposes the construction of data centers.
“While other organizations are asking how to use AI responsibly, Greenpeace is questioning whether it is even necessary to build all these massive data centers in the first place,” says Katrin Xueying Wu, a climate and energy campaigner at the organization.
Other groups, meanwhile, now view AI as essential for effective environmental protection efforts.
“It is our responsibility to save the planet. If we can use AI to make more effective and efficient use of our limited resources to achieve this, then we must seriously consider that possibility,” says Stuart Butchart, chief scientist at BirdLife International, a UK-based organization.
“If we don’t, we would essentially be failing in our duty.”