WWF-Singapore (World Wide Fund for Nature Singapore)

Identified Problem

Illegal wildlife trade is the second largest driver of species extinction, following habitat loss. Identifying illegal wildlife products from online marketplaces is challenging and time-consuming. There is a need to reduce the turnaround time between the collection of data and sharing of validated data with stakeholders to effectively combat the illegal trade.

AI Solution

A machine learning solution was developed to analyze online listings of illegal wildlife products. The AI model uses both text and images from these online listings to make predictions. This solution is intended to reduce the turnaround time in the identification of illegal wildlife products.

Key Features

• The AI model uses both text and images from online listings to make predictions.
• It can identify the data class that a particular listing belongs to.
• The model was trained using data from the Cyber spotters program, which included information on elephant ivory, wildcat products, and pangolin scales.
• The model has been adapted to incorporate a translator and to include emojis from online listings.
• Additional data on sea turtles was incorporated during the model development.

Benefits of AI Solution

• The AI solution reduces the turnaround time between data collection and sharing with stakeholders by reducing the need for manual verification.
• It is able to give a significantly higher number of accurate predictions on the listings.
• It is a key solution to ensuring that the issue of illegal wildlife trade can be addressed in a timely manner.

Impact of AI Solution

• The project demonstrates the importance of collaboration between people who are passionate about wildlife and tech people to combat illegal trade.
• The AI model is in the testing phase and there are plans to make it widely used in other WWF offices.
• Complementary systems are being developed to further automate the process of data collection and sharing.