Periodic Reporting for period 3 - WILDTRADE (Quantifying the global patterns and trends of the illegal wildlife trade: from artificial intelligence to financial market analysis)
Período documentado: 2022-06-01 hasta 2023-11-30
The illegal wildlife trade in one of the major threats affecting the global biodiversity crisis. Wildlife trade is also linked to the spreading of zoonotic diseases, such as SARS-CoV-2. International and European policies call for actions to halt illegal wildlife trade. In the Digital Age, an important part of the wildlife trade has moved to online platforms, especially social media platforms. The WILDTRADE project is using machine learning methods and novel data sources mined from social media and other digital platforms, in combination with newly collected economic data, to quantify global patterns and trends of the illegal wildlife trade and how market forces shape them.
We have developed novel application methods to automatically collect and analyze textual, visual, and audio content from multiple digital platforms. An end-to-end pipeline starts from searching and downloading information about species threatened by wildlife trade and proceeds with implementing natural language processing and machine learning methods to filter and retain only relevant information for further analyses. A ‘Named Entity Recognition model’ is being used to extract additional relevant information. This information includes reported prices and quantities of traded animals. The data collection framework follows data privacy and protection safeguards to comply with the European Union's General Data Protection Regulation.
The resulting database is currently being used to identify the main trade routes of the illegal wildlife trade and map the movement of wildlife products. We have already examined the spatial characteristics of the songbird trade using multiple online data sources, including citizen science data, small advertisements from online marketplace platforms, and videos. In this case, it was confirmed that data from digital sources can give rich insights into the spatial, temporal and taxonomic structure of wildlife trade. We have also used online surveys and choice experiment techniques to investigate how the perception of rarity is driving consumers’ preferences for exotic pets using an online survey. A novel framework to summarize and generate insights on demand for rarity and scarcity in the wildlife trade was also developed. We further proposed that the poaching economy of wildlife products such as elephant ivory and rhino horn is driven by a combination of persistent consumer demand and market speculation, and enabled by weak governance, lack of adequate resources for species protection, and alienation of local people who pay the costs of living alongside these species. Strategies that move toward empowering local communities with stronger property rights over wildlife and delivering more benefits to them, are underused elements in the current fight against the onslaught of the international illegal wildlife trade.
The resulting database is currently being used to identify the main trade routes of the illegal wildlife trade and map the movement of wildlife products. We have already examined the spatial characteristics of the songbird trade using multiple online data sources, including citizen science data, small advertisements from online marketplace platforms, and videos. In this case, it was confirmed that data from digital sources can give rich insights into the spatial, temporal and taxonomic structure of wildlife trade. We have also used online surveys and choice experiment techniques to investigate how the perception of rarity is driving consumers’ preferences for exotic pets using an online survey. A novel framework to summarize and generate insights on demand for rarity and scarcity in the wildlife trade was also developed. We further proposed that the poaching economy of wildlife products such as elephant ivory and rhino horn is driven by a combination of persistent consumer demand and market speculation, and enabled by weak governance, lack of adequate resources for species protection, and alienation of local people who pay the costs of living alongside these species. Strategies that move toward empowering local communities with stronger property rights over wildlife and delivering more benefits to them, are underused elements in the current fight against the onslaught of the international illegal wildlife trade.
The WILDTRADE project has successfully developed tools for efficient monitoring of the illegal wildlife trade on social media platforms and assessed how market forces shape the trade. Our application of deep neural networks for automated content identification in the context of the illegal wildlife trade are especially innovative. The remaining work in the second part of the project will focus on identifying areas of high pressure from wildlife trade and understanding the patterns of supply and demand from supply to demand countries. We expect to carry out analyses at multiple scales, from global to local, combining multiple data sources, including digital data sources, to address these questions.