Automated monitoring of websites that trade wildlife is increasingly necessary to inform conservation and biosecurity efforts. However, e-commerce and wildlife trading websites can contain a vast number of advertisements, an unknown proportion of which may be irrelevant to researchers and practitioners. Given that many wildlife-trade advertisements have an unstructured text format, automated identification of relevant listings has not traditionally been possible, nor attempted. Other scientific disciplines have solved similar problems using machine learning and natural language processing models, such as text classifiers. Here, we test the ability of a suite of text classifiers to extract relevant advertisements from wildlife trade occurring on the Internet. We collected data from an Australian classifieds website where people can post advertisements of their pet birds (n = 16.5k advertisements). We found that text classifiers can predict, with a high degree of accuracy, which listings are relevant (ROC AUC ≥ 0.98, F1 score ≥ 0.77). Furthermore, in an attempt to answer the question ‘how much data is required to have an adequately performing model?’, we conducted a sensitivity analysis by simulating decreases in sample sizes to measure the subsequent change in model performance. From our sensitivity analysis, we found that text classifiers required a minimum sample size of 33% (c. 5.5k listings) to accurately identify relevant listings (for our dataset), providing a reference point for future applications of this sort. Our results suggest that text classification is a viable tool that can be applied to the online trade of wildlife to reduce time dedicated to data cleaning. However, the success of text classifiers will vary depending on the advertisements and websites, and will therefore be context dependent. Further work to integrate other machine learning tools, such as image classification, may provide better predictive abilities in the context of streamlining data processing for wildlife trade related online data.

No abstract

The trade and keeping of exotic pets has serious implications for both biosecurity and biodiversity conservation. In Australia, the online trade of live invertebrates is an understudied and unregulated issue, with almost non‐existent monitoring. It is uncertain what species are being traded, whether they are being identified correctly, and how they are being sourced (i.e., captive bred or wild harvested, native, or alien). Consequently, potential invasion risks and conservation concerns remain unknown. Here, we explored the online trade of terrestrial invertebrates in Australia across a range of publicly available e‐commerce platforms. We detected 264 species of invertebrate traded, from 71 families and 168 genera over 12 months. The native Extatosoma tiaratum (giant prickly stick insect) was the most traded species, while the most popular families were Phasmatidae (stick insects), Formicidae (ants) and Theraphosidae (tarantulas). Three species are known to be invasive in Australia, while 87% of species traded were native. The conservation status of almost of the species (92%) listed in the invertebrate trade has not been evaluated. Exploring socio‐demographic relationships, we found that human population density was positively correlated with the location of invertebrate sellers. Further, we found the classifieds website had lower prices in contrast to traditional online pet‐stores (median of c. A$7 less). Finally, we did not observe a saturation in the number of species traded in our one‐year study, exemplifying the need for large scale monitoring and risk assessments for Australia's online terrestrial invertebrate trade. We recommend continued surveillance of live invertebrate trade on e‐commerce sites. Substantial changes to legislation and monitoring methods are required at a national level to control the vast number of invertebrates traded across the country, and to minimise the future risks of the invertebrate trade.

Contemporary wildlife trade is massively facilitated by the Internet. By design, the dark web is one layer of the Internet that is difficult to monitor and continues to lack thorough investigation. Here, we accessed a comprehensive database of dark web marketplaces to search across c. 2 million dark web advertisements over 5 years using c. 7 k wildlife trade‐related search terms. We found 153 species traded in 3332 advertisements (c. 600 advertisements per year). We characterized a highly specialized wildlife trade market, where c. 90% of dark‐web wildlife advertisements were for recreational drugs. We verified that 68 species contained chemicals with drug properties. Species advertised as drugs mostly comprised of plant species, however, fungi and animals were also traded as drugs. Most species with drug properties were psychedelics (45 species), including one genera of fungi, Psilocybe, with 19 species traded on the dark web. The native distribution of plants with drug properties were clustered in Central and South America. A smaller proportion of trade was for purported medicinal properties of wildlife, clothing, decoration, and as pets. Synthesis and applications. Our results greatly expand on what wildlife species are currently traded on the dark web and provide a baseline to track future changes. Given the low number of advertisements, we assume current conservation and biosecurity risks of the dark web are low. While wildlife trade is rampant on other layers of the Internet, particularly on e‐commerce and social media sites, trade on the dark web may still increase if these popular platforms are rendered less accessible to traders (e.g., via an increase in enforcement). We recommend focussing on surveillance of e‐commerce and social media sites, but we encourage continued monitoring of the dark web periodically to evaluate potential shifts in wildlife trade across this more occluded layer of the Internet. Read the free Plain Language Summary for this article on the Journal blog.

The illegal wildlife trade (IWT) threatens conservation and biosecurity efforts. The Internet has greatly facilitated the trade of wildlife, and researchers have increasingly examined the Internet to uncover illegal trade. However, most efforts to locate illegal trade on the Internet are targeted to one or few taxa or products. Large-scale efforts to find illegal wildlife on the Internet (e-commerce, social media, dark web) may be facilitated by a systematic compilation of illegally traded wildlife taxa and their uses. Here, we provide such a dataset. We used seizure records from three global wildlife trade databases to compile the identity of seized taxa along with their intended usage (i.e., use-type). Our dataset includes c. 4.9k distinct taxa representing c. 3.3k species and contains c. 11k taxa-use combinations from 110 unique use-types. Further, we acquired over 45k common names for seized taxa from over 100 languages. Our dataset can be used to conduct large-scale broad searches of the Internet to find illegally traded wildlife. Further, our dataset can be filtered for more targeted searches of specific taxa or derived products.