Stewards Magazine

Although artificial intelligence (AI) programs were first developed decades ago, they made an undeniable splash in 2023. Now, researchers at the University of Michigan School for Environment and Sustainability (SEAS), alongside collaborators in the U-M School of Information, are focused on laying the groundwork for using AI to accelerate wildlife conservation efforts using social media.

Nathan Fox, a SEAS and Michigan Institute for Data Science postdoctoral fellow, is an environmental scientist who is one of 10 researchers in the 2023 cohort of the Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship. Fox is working with SEAS Assistant Professor Derek Van Berkel and Associate Professor Neil Carter to use geolocated imagery from social media to complement and enhance existing conservation datasets.

“We want to fill in the blanks in datasets that we already have. Maybe we’ll find images of a species outside of the range of where we knew it to be, and then we can start to understand why they moved, say, further north, because of climate change,” says Fox. “This could lead us to focus conservation efforts in a different area since the information we’re gaining from social media shows a different picture than what traditional methods were showing.”

In traditional conservation science, maps are used to track where animals are located, but with social media and AI, the research team uses spatial modeling to predict where habitats are or where they might be moving. As part of Fox’s two-year fellowship, he is training computer vision models to pick up on specific wildlife characteristics and provide probability guesses. Carter says this work will contribute to wildlife conservation in a way that’s never been done before.

“This is a brand-new scientific endeavor where we’re at the early stages of defining the problem and refining an approach to tackle that problem. We have a sense that there are potentially transformative areas here,” says Carter. “So this work can bring new tools, potentially extremely powerful ones, to bear on a set of growing sustainability challenges.”

Van Berkel adds that this work would be challenging without the use of AI. “Early on, we were looking at photographs manually. So, if you have a set of 50 million photos, you can only look at a sample of say, 500,” says Van Berkel. “As with anything else, there are still some limitations, such as what types of wildlife are often excluded from imagery posted on social media, but with AI, and computer vision in particular, we have an amazing opportunity to use all of the readily available data to map and identify trends in wildlife and watch how they change over time due to climate change and other circumstances.”

The Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship is part of Schmidt Futures, a global partnership that aims to apply artificial intelligence to research in science, technology, engineering and mathematics in an effort to accelerate the next scientific revolution. U-M is one of nine universities selected to join the fellowship program. Bill Currie, SEAS professor and associate dean for research and engagement, is co-principal investigator on the Schmidt Futures award and co-director of the fellowship program.