When you think of poaching, what comes to mind? Do you think of the last white rhino on earth being protected 24/7 by a foray of armed guards? Maybe a Yosemite Sam-type character roving through the savanna with an elephant gun looking to score the pelt of a big cat? You probably don’t consider rangers using artificial intelligence to outsmart poachers. PAWS, an innovative and fascinating way to protect animals, outwits poachers and protects wildlife rangers on the frontlines of poaching with AI technology.
PAWS stands for Protection Assistant for Wildlife Security, created in 2013 by 2013 by Milind Tambe, a professor of computer science at the University of Southern California, and a team of Ph.D. students, one of them being Fei Fang, who came up with the concept of applying game theory and artificial intelligence to conservation. Fei Fang, an assistant professor at Carnegie Melon’s Institute for Software Research, has worked on applying AI to sustainability, security, and conservation throughout her academic and professional tenure. In a field primarily dominated by men, Fang stands out as a leader in software and conservation. Like the rest of the world, Tambe, Fei, and their team were alarmed by the rate at which savvy poachers were slaughtering animals like tigers and elephants. Poachers were learning patrol routes, using GPS technology, night vision goggles, and other high-tech products that allowed them to slip in and out of nature reserves like ghosts. PAWS is an attempt to conserve endangered wildlife while beating poachers at their own game through the use of machine learning to study where poachers are striking, then creating randomized patrol routes based on where poachers are expected to strike next.
PAWS is based on algorithms borrowed from the Department of Homeland Security and the Transportation Security Administration, which were designed to stop smuggling and trafficking at America’s entry points and borders. Tambe and his team first applied PAWS in Uganda’s Queen Elizabeth National Park, where the algorithm’s dataset was built over 14 years of animal observation, poaching activity, and other environmental observations. After adjusting for terrain and the human element of the patrol route nighttime navigation, PAWS’ algorithm eventually learned to create randomized routes that outperformed routes not generated by PAWS, with rangers able to see more human and animal activity. As time went on, PAWS got smarter as it learned and acquired more data to predict poacher activity.
These randomized routes held several essential benefits. As poachers cannot learn patrol routes or predict where they might be, the risk of being caught increases, thus deterring activity. PAWS also protects wildlife rangers, who have also been known to be targeted by poachers to reach their protected animals. As wildlife are not the only target for poachers, solutions should be built with rangers in mind. With both human and animal lives at stake, conservation and protection are vital sectors to improve. A key element of PAWS’ method involves going beyond creating a new solution to protect wildlife by also ensuring the safety of the people tasked with protecting endangered animals. In this context, the use of AI makes humans and wildlife smarter and safer.
References
Fang, Fei, Thanh H. Nguyen, Rob Pickles, Wai Y. Lam, Gopalasamy R. Clements, Bo An, Amandeep Singh, Milind Tambe, and Andrew Lemieux. “Deploying PAWS: Field optimization of the protection assistant for wildlife security.” In Twenty-Eighth IAAI Conference. 2016; Khalid, Samia. (2018, July 23). – AI, Drones and Game Theory for Smart Wildlife Conservation. – Medium. https://medium.com/datadriveninvestor/ai-drones-and-game-theory-for-smart-wildlife-conservation-7c1a27adc151
Snow, Jackie. (2016, June 12). – Rangers Use Artificial Intelligence to Fight Poachers. – National Geographic. https://www.nationalgeographic.com/news/2016/06/paws-artificial-intelligence-fights-poaching-ranger-patrols-wildlife-conservation/
Photo by Courtenay Crane