CALI, COLOMBIA— In 1992, when more than 150 nations agreed to the Convention on Biodiversity (CBD), a landmark global pact aimed at protecting the environment, the iPhone didn’t exist, the modern internet was in its infancy, and researchers trying to develop artificial intelligence (AI) were suffering through a period of scientific and financial setbacks that became known as the “AI winter.”

How times have changed. This week, as an estimated 15,000 people gather here in Colombia for the 16th meeting of the CBD’s parties (or COP16), iPhones and the internet are taken for granted—and AI has emerged as a potentially pivotal tool for achieving an ambitious goal of halting and reversing biodiversity loss by 2030. Researchers at the meeting are predicting AI tools will have a growing impact on conservation, from accelerating efforts to comb through vast troves of field data to supporting transparent, fast ways to finance efforts to protect endangered species.

“The advances in AI technology can help us to understand much better what we have in terms of biodiversity, understand what state it is in, and how we are losing it,” says Ane Alencar, science director at the Amazon Environmental Research Institute in Brazil.

One issue getting attention is how AI can help the 196 nations that ratified the CBD meet their commitment to survey and monitor their biodiversity. Traditionally, that could only be done through labor-intensive fieldwork aimed at finding and tallying species and identifying rare organisms or vulnerable habitats in need of protection. But researchers increasingly use automated digital tools, including audio recorders, camera traps, and satellite sensors, to monitor plants and animals. And AI tools can be much faster and more efficient than humans at sifting through hundreds of hours of sound or video recordings to identify species of interest.

AI can reduce the time needed to complete such tasks “from years to minutes,” says Tanya Berger-Wolf, a computer science professor at Ohio State University and co-founder of the AI conservation software nonprofit Wild Me. And that can mean big time savings for governments and conservation groups that have limited staff and money, she adds. In a 2022 paper published in Global Ecology and Conservation, for example, researchers found an AI application increased the processing speed of camera trap data by more than 500%.

At one packed session here, researchers discussed a competition aimed at further accelerating field surveys using AI. In July, six teams of scientists, including a team from Brazil, vied for the $10 million Rainforest X prize, which challenges researchers to find the best way to survey 100 hectares of Brazilian rainforest in just 24 hours, then produce the most useful insights within 48 hours. The competing researchers noted that AI did the heavy lifting of analyzing data gathered via drones, audio sensors, and analyses of environmental DNA, all the while respecting Brazilian data collection rules. The winner won’t be announced until next month, but session organizers say the effort highlights how AI-powered projects could inform management and policy decisions in near real time, not the years it now takes.

“The intention behind the prize was to get to the point where [AI] can be accessible and affordable,” says Peter Houlihan, executive vice president of biodiversity and conservation at the XPRIZE Foundation. The group also wants local and Indigenous users to be able set up AI systems without having to take “technical deep dives.”

AI could also help support emerging funding mechanisms for conservation projects, says economist Anna Ducros of the International Institute for Environment and Development. For example, camera traps in the Amazon could, hypothetically, automatically alert investors who helped finance a conservation project when they spot a particular kind of animal, such as a jaguar. “Analyzing biodiversity data more accurately and faster opens up space for this data to be passed on to different stakeholders,” she says.

While the pressure mounts on negotiators at the meeting to reach a new agreement on a range of issues, including how to fund global biodiversity protection efforts, AI’s use in conservation continues to grow quietly. The meeting is set to end on Friday.

Source: science

Une réaction ?
0Cool0Bad0Lol0Sad