Finbarr O'Reilly / The New York Times

The ELP tracks various elephant herds in Africa through the use of sensors that detect patterns of elephant calls and gunshots.

September 6, 2018

Elephant Listening Project Uses Artificial Intelligence to Combat Poaching in Africa

Print More

Cornell University’s Elephant Listening Project has been using artificial intelligence technology to track African forest elephants to safeguard them from poachers and other threats.

The ELP tracks various elephant herds throughout the 580-square-mile Nouabalé-Ndoki National Park, located in the Republic of Congo, through the use of 50 sensors placed in various locations in the park.

The sensors were developed by an AI startup called Conservation Metrics. The sensors track patterns of elephant calls and gunshots and report the information to park rangers. The rangers then use this information to determine where the herds are gathered and where the poachers are operating.

The ELP, which is part of the Cornell Lab of Ornithology, is hoping to reverse the impact poachers have had on the forest elephant population in central Africa. The Daily Mail found that, in 2014 alone, 25,000 forest elephants were killed by poachers in Minkébé National Park in Gabon.

However, filtering through the audio recordings to find the relevant information is no small task — the sensors generate over seven terabytes of recordings every three months, which is equivalent to just over 2 million songs on iTunes.

The Elephant Listening Project hopes to speed up the process and thus save more elephants from poaching through the use of Conservation Metric’s AI software.

“A key thing this collaboration will do is speed things up, so we can show the people who manage the national park that we can provide information that will make a difference,” Peter Wrege Ph.D. ’80, director of the Elephant Listening Project, told The Daily Mail. “If it takes us a year to figure out what elephants are doing in the forest, it’s already too late.”

Before the use of Conservation Metric’s software, it would take ELP six to eight weeks to run all of the sound recordings through its various computers. The recordings then had to be screened by hand to test the accuracy of the computers.

The entire process usually took about three months from start to finish. Now, through the use of AI, the process has been shortened to just 22 days and could potentially become even faster in the future.

“Acoustics isn’t going to stop the poaching, but I do think it offers maybe the only way we can get information regularly enough,” Wrege told The Daily Mail. “It’s daunting, but it’s worth it, and it can be done. We just have to keep at it.”