Environmental groups worldwide are using artificial intelligence to analyse climate data, track forest fires and count endangered wildlife - embracing a technology they also warn is driving a surge in electricity demand, water use and greenhouse-gas emissions. The contradiction is forcing these organisations to justify their AI use in carrying out missions to combat climate change or protect natural resources.
"We recognise that AI can come with environmental costs, and we need to be thoughtful about how we use it," said Ali Swanson, AI lead at the US nonprofit Conservation International. "But at a time when the climate and biodiversity crises are becoming more urgent, AI can help us move faster and make better use of the information we're collecting."
The group recently started using AI to assess hotspots detected by satellites and send real-time alerts to community forest patrols about potential wildfires in remote and environmentally sensitive parts of Cambodia. All said they used the technology in some form and expressed concern about the environmental impact of the data centres powering it.
The centres, filled with servers, can require vast amounts of electricity, along with water for cooling. It's a global dilemma for individuals, governments and industries balancing AI's benefits with its enormous risks.
The tradeoffs likely will grab attention at the UN's COP31 climate conference this month, when host Turkey will call on governments, businesses and other institutions to sign a pledge outlining principles to ensure AI is developed and deployed sustainably. UN Secretary-General Antonio Guterres wants major AI companies to publicly disclose their systems' environmental impacts.
Researchers at the United Nations University Institute for Water, Environment and Health estimate the global data-centre footprint could more than double by 2030, reaching a size about twice that of Jakarta, the Indonesian megalopolis. The International Energy Agency, meanwhile, estimates global data centres will consume about 945 terawatt-hours of electricity annually by 2030, producing the equivalent of more than 300 million metric tons of carbon dioxide.
That's double current consumption and about equal to Japan's annual electricity use. The agency also projects that global data centre water consumption will increase from about 560 billion litres (147.9 billion gallons) in 2023 to about 1.2 trillion litres (317 billion gallons) annually by 2030.
Technology giants including Microsoft, Amazon and Alphabet's Google have reported rising emissions linked in part to data centre expansion, raising questions about their ability to meet their own long-term climate goals. Amazon, Microsoft and Google didn't respond to requests for comment.
Many of the organisations that spoke to Save the Elephants, a partner of the Wildlife Conservation Network, said it is using machine learning to improve population surveys and study elephant communication. Frank Pope, the group's chief executive, said an aerial survey of Kenya's largest protected area last year generated roughly 200,000 images that took nine people nine months to review manually.
AI models trained on Nvidia's GeForce graphics platforms can now complete similar analyses in hours. "The results have been eye-opening," Pope said.
"We found that human surveys were missing a significant portion."