Using AI to investigate the illegal wildlife trade
Each year, millions of wild animals are sold illegally into a global trade worth billions of dollars annually, as pets, trophies or to supply traditional medicines – among other uses. The impacts span across the tree of life, affecting fauna, flora, and fungi. Since the advent of the internet and social media, the illegal wildlife trade has been booming. Yet there is a lack of data able to show the exact volumes of trade, and the availability of products on the market. “Estimating the scale of the illegal wildlife trade is challenging given the clandestine nature of wildlife crime,” explains Enrico Di Minin, professor of conservation geography at the University of Helsinki. “The wildlife trade can be legal or illegal at various stages of the trade chain and it can be hard to differentiate,” he notes. In the EU-funded WILDTRADE project, Di Minin led a team of researchers investigating the global patterns and trends in the illegal wildlife trade. The project used automated data mining and analysis techniques to seek out the trade on social media and other digital platforms, to gain a clearer picture of its extent.
Data mining from digital media
The project developed novel application methods to automatically collect and analyse textual, visual, and metadata content from social media and other digital platforms. The WILDTRADE system first searches and downloads information about species threatened by wildlife via an API, then uses natural language processing, computer vision, and multimodal learning methods to filter and retain only relevant posts and information, which are then used in follow-up analyses. Using their new methodology, the researchers were able to identify many animal, plant, and fungi species and products for sale across multiple social media and other digital platforms, many of which were threatened species.
Understanding market drivers behind the global trade
While the extent of online trade appears to be global in nature, the team found that the species traded, and the digital platforms involved, appear to be context dependent. “Our findings indicate that the different stages of the trade chain may be geographically distinct, and extensive transport networks and commercial captive breeding facilities may play significant roles in the wildlife trade,” he says. The findings also suggested that limiting the legal supply of wildlife products has fostered conditions that allow the poaching economy, or ‘poachernomics’, to flourish. “We have proposed that empowering local communities through strengthened property rights and increased benefits may be an effective strategy to combat wildlife crime,” he adds. However, the researchers did find that emotions such as attachment, affection and nurturing instincts were strong motivators among exotic pet owners, who showed a willingness to support species conservation in the wild. “This underscores how relational factors can positively impact conservation efforts,” says Di Minin.
Helping conservation efforts
The methods developed in WILDTRADE could be repurposed to investigate the links between online wildlife trade and zoonotic diseases and could also feed into mobile applications for conservation organisations to monitor online wildlife trade more effectively. “By harnessing social media big data, for example as part of the Global Biodiversity Framework and the Convention on International Trade in Endangered Species of Wild Fauna and Flora, the results can be used as indicators of (un)sustainable wildlife use and trade,” says Di Minin.
Keywords
WILDTRADE, illegal, wildlife, trade, analysis, data mining, digital media, market drivers, global
