The boom in artificial intelligence, or AI, is transforming many aspects of research, engineering and design, including work at the forefront of climate action. AI and machine learning are already aiding environmental science and the development of more sustainable products and clean energy options.
One recent example from researchers at Google Maps and Google Earth shows how AI has become an important tool to improve monitoring of changes in the environment. The AI-powered system called Cloud Score+ is providing clearer, more detailed images of remote ecosystems that are frequently covered by clouds.
“When you’re trying to get a clear view of the Earth’s surface from space, clouds (and their shadows) can get in the way,” Google Geo Sustainability Research Scientist Valerie Pasquarella told Newsweek via email.
Courtesy of Google Research
Previous efforts to automate the detection of clouds and shadows in satellite imagery proved ineffective, she said, because their “one-size-fits-all” approach didn’t work for challenges such as the high, thin cirrus clouds or faint shadows on the landscape.
“It gets particularly challenging when you’re in a cloudy region of the world where any relatively clear appearance is better than nothing,” Pasquarella said.
Some of the world’s most imperiled and important tropical forest ecosystems fall into that category. The rainforests and mountainous cloud forests in the world’s major tropical forest basins in Africa, Asia and South America store massive amounts of carbon dioxide that would otherwise contribute to further warming in the atmosphere, and the forests are of course home to astonishing biodiversity. But they also face threats from illegal logging, mining and ranching and the increasing effects of climate change. Ecologists, land stewards and forest managers need a clear view of what is happening to counter those threats.
Pasquarella explained that Cloud Score+ uses video AI to score every pixel of every satellite image.
“We can pick and choose only the highest quality pixels for visualization or analysis,” she said.
Courtesy of Google Research
The results are already being put to use in some of the world’s cloudiest places, including the mountains of Ecuador. Imbabura Province in the country’s northern Andes was designated a Global Geopark by UNESCO in 2019. But constant clouds and remote terrain led to a poor understanding of the region’s assets.
Google researchers worked with the local government to offer improved images of the region generated with Cloud Score+.
“Cloud Score+ has also been used to help identify high-quality, cloud-free pixels for monitoring forest health and disease outbreaks, detecting plant stress in peatlands, and mapping forest tree species for improved management,” Pasquarella said.
Courtesy of Google Research
Google Research projects are also applying AI to better monitoring and detecting a broad range of climate impacts in developed areas, Pasquarella said, including improvements in wildfire detection, enhanced flood prediction and a heat resilience tool to help cities address extreme temperatures.
“We believe that AI can play a transformative role in addressing the effects of climate change,” she said.
Google is also, however, an example of the climate damage that comes from AI. Google parent company Alphabet appears on Newsweek‘s ranking of America’s Greenest Companies, where it gets 4 out of 5 stars, and has ambitious climate targets to limit its greenhouse gas emissions. However, Google’s most recent sustainability report shows a sharp jump in emissions, largely due to the buildout of power-hungry data centers behind the AI boom.
Courtesy of Google Research
A Newsweek Horizons event last month during Climate Week NYC addressed the dual nature of AI’s impact. In the panel discussion “AI: Climate Hero or Climate Villain?” experts from Big Tech, energy and academia explained how AI offers some climate solutions while creating an energy problem of its own.
Panelist Priya Donti is an MIT assistant professor and cofounder of Climate Change AI. That collection of scientists researching AI’s effects on climate action is also finding new ways for AI to help address and adapt to climate impacts. Donti pointed to a project that helps people of Fiji better predict flooding events on the island and AI tools for developing materials important for clean energy technology.
Marleen Moise
However, Donti said, AI is also used in the exploitation of fossil fuels and to encourage mass consumptions of goods, hardly the stuff of environmental progress.
Ultimately, she said, AI is not what will make the decisions for or against climate action, “people and organizations are,” and AI is just a tool—albeit a powerful one.
“AI is an accelerator of what we do across society,” Donti told the audience at Newsweek‘s New York office. “If we’re serious about fighting climate change, it will be a hero. If we’re not, it’s going to be a villain.”