Environmental groups embrace AI, even as they warn of data center impacts

Environmental groups worldwide are harnessing artificial intelligence (AI) to analyze climate data, monitor forest fires, and count endangered wildlife. However, this embrace of AI is also accompanied by warnings about its potential to increase electricity demand, water usage, and greenhouse-gas emissions. Organizations are grappling with the contradiction, striving to justify their use of AI while acknowledging its environmental costs.

Ali Swanson, AI lead at Conservation International, recognizes these costs but emphasizes that AI can expedite efforts to combat climate change and protect natural resources. The group recently employed AI to identify hotspots from satellite data and alert community forest patrols in Cambodia about potential wildfires. A Reuters survey of 15 major environmental and conservation organizations revealed that all utilize AI in some form, expressing concerns about the environmental impact of data centers that power this technology.

Data centers, housing servers, consume substantial electricity and require water for cooling. As the global data-center footprint is projected to nearly double by 2030, equal in size to twice Jakarta, the International Energy Agency estimates it will consume about 945 terawatt-hours of electricity annually, producing over 300 million metric tons of carbon dioxide – a figure equivalent to Japan's annual electricity use.

Moreover, global data center water consumption is expected to increase from 560 billion liters in 2023 to 1.2 trillion liters by 2030. Despite these concerns, many organizations believe AI's benefits outweigh its costs because the technology enables tasks that were previously difficult, expensive, or impossible. For instance, Kenya's Save the Elephants uses machine learning to enhance population surveys and study elephant communication, while the Jane Goodall Institute leverages AI to digitize decades of field notes on chimpanzee behavior in Tanzania's Gombe National Park, reducing the time required from 13-15 years to about two.

Some groups experiment with limiting the environmental impact of AI use, such as the World Resources Institute and the Nature Conservancy, which optimize resource-intensive applications and process information aboard vessels instead of uploading videos to centralized data centers.