Updated: Jun 11, 2019
The same software that is used on social media platforms to recognize your face in photos is now being used to track animals across the illegal wildlife trade. Still in the early stages of development, this new AI conservationist tool uses photos of chimpanzees to perfect its facial recognition abilities and help locate animals as they pass through online wildlife markets.
By searching across social media, the algorithm uses this training to locate trafficked animals and bring them to the attention of authorities. Once an animal is identified, authorities are informed of the account and can begin to track the account holder, potentially preventing more animals from falling into the trade.
The criminal networks supporting these trades use social media to secretly buy and sell animals across country lines and have been notoriously difficult to stop. For some species, this pervasive trade threatens their safety and puts them within easy reach of those looking for an exotic pet or performer in a commercial zoo as a baby chimp bought online costs an average of $12,500.
Social Media’s Impact on Illegal Trafficking
Conservation Biologist Alexandra Russo spearheaded the project, aptly titled ChimpFace. As she investigated the online chimpanzee marketplace, she began to notice how social media sites like Instagram and Facebook unwittingly contributed to the trade. Sellers would use the sites to post photos of the animals in their possession and buyers would use them as a way to connect with these sellers. But even though conservation organizations were aware of the problem, a sustainable solution was difficult to find.
"We were spending more and more time looking through the depths of the internet,” said Russon, “it's like a rabbit hole. You don't know where to look. You click around pretty aimlessly until you start to find things that look suspicious. So, I thought there must be a more efficient way to do this.
"I began discussing the possibility of using some kind of software that could automatically find ape faces in online searches."
With the desire to do more to end the trade, Russo contacted Conservation X Labs, a Washington D.C. based non-profit group dedicated to creating technological solutions to modern conservation challenges. Computer vision and AI expert Dr. Colin McCormick soon joined the project, and from their combined efforts, ChimpFace algorithm was created.
Development and Training of ChimpFace Algorithm
ChimpFace is an AI algorithm that is trained to identify unique chimpanzee faces in images across social media platforms. The program begins by determining that a chimpanzee is in the photo and then identifying the unique characteristics of that individual chimp.
"We train the algorithm,” said McCormick, “to recognize individual chimps using up to 30 photos of an individual, adapting it to the chimp's facial structures."
The initial process takes less than a second, but as more photos are added, processing speed decreases. However, these extra photos are essential to the success of the project as the algorithm uses them to understand how factors like lighting, posture, and facial expressions affect the image and the chimp.
Currently operating with over 3,000 chimp photos of chimp faces, the algorithm draws a square around the face in the photo and cross-examines it against the photos in its database to find a match. As word of the project has spread, other organizations such as Brazil’s Great Ape Project and Guinea’s Project Primate International have contributed their own pictures of both wild and sanctuary based chimps to the database.
"They send us images of known named apes so we can build the individual recognition algorithm,” said McCormick, but the fact that chimpanzee faces are so close to human faces creates a harder technical challenge.”
Since humans share an estimated 98% of our genes with humans, determining what’s a human and what’s a chimp can be especially difficult for AI technologies and is one of the key challenges for ChimpFace.
Future Growth of Facial Recognition Technology
The team behind ChimpFace hope to continue perfecting the technology and to eventually provide a streamlined platform for local police and other authorities to use themselves. With over 2,000 chimpanzees pulled from the wild every year to be bought and sold in the trafficking market, demand for this tool is high amongst the countries most affected.
"We want to integrate the tool into other rescue organizations to then scale up and out so we can start to really wrap our heads around these large-scale trade networks," said Russo. "...We're looking to partner with social media companies who could incorporate ChimpFace into their crime monitoring. Then, we could put a big dent in wildlife trafficking.”
The success of this project could have wide-reaching benefits for other illegally trafficked species. If the algorithm is capable of successfully and efficiently identifying chimps, it may also be able to identify trafficked and endangered tigers, pangolins, leopards, and other great apes and help bring an end to the online marketplace for these products.