AI Tool Exposes Hidden Networks of Online Wildlife Trafficking
A new system has been developed to track and disrupt online wildlife trafficking. Created by Dr. Juliana Freire, it scans millions of advertisements across multiple platforms to identify illegal sales. The tool aims to expose hidden trade networks and support policy changes that could curb this growing problem.
Wildlife trafficking remains one of the world's largest illegal industries. It threatens biodiversity, funds organised crime, and poses serious public health risks. Now, a significant portion of this trade has shifted online, blending with everyday e-commerce listings.
Dr. Freire's team analysed 23 major online marketplaces for wildlife products. Their system gathered over 11 million advertisements in just over a month. Using large language models, it labels data and builds specialised classifiers to detect different species, product types, and platforms.
The tool automatically extracts key details from listings, including titles, descriptions, prices, images, and seller information. This helps identify trends and high-risk areas. For example, reptile leather products—such as crocodile, alligator, and python skins—dominated eBay listings, with just ten product combinations making up 72% of all sales.
Geographic patterns also emerged. While small reptile leather items were shipped from 65 countries, 90% of listings came from only ten nations. Even banned items, like tiger shark trophies, still appeared on eBay despite existing restrictions.
The research suggests that targeting key sellers or product types could disrupt large parts of the wildlife market. Simple policy adjustments, guided by this data, may help reduce illegal trade more effectively.
The findings reveal how online wildlife trafficking operates across borders and platforms. By pinpointing high-volume products and sellers, authorities could take more focused action. The system provides a clearer picture of the trade, offering a way to tackle it with data-driven policies.
