Cyber ​​Spotters Fight Wildlife Smuggling

SINGAPORE (ANN/THE STRAITS TIMES) – Exposing the illegal wildlife trade thriving on e-commerce platforms is a relentless game of whack-a-mole, made even more difficult by the use of artificial intelligence (AI) tools.

According to Jayasri Srikantan, head of the volunteer initiative at the World Wildlife Fund (WWF), online listings for items such as ivory, pangolin scales and live animals often use emoticons, obscure language and coded terms such as “rare collectibles”, making it difficult to detect finds.

However, an artificial intelligence model is being developed to help WWF volunteers – known as cyberspotters – more accurately evaluate flagged listings on Shopee, Facebook Marketplace and other platforms.

Trained on thousands of samples of wild produce and their estimated black market prices, the system helps prevent false positives, strengthening the group’s credibility with law enforcement partners.

The software’s accuracy is currently around 90 per cent, Ms Jayasri said, adding that new code words and trends are added to the system regularly to keep it up to date with current trends.

WWF’s Cyber ​​Spotter AI model, which will be rolled out in 2023, is part of an initiative to fund and launch AI projects in Singapore, with the aim of driving AI adoption. The initiative is led by the national AI Singapore programme.

The AI ​​system, which has been in the works since 2022, aims to improve the accuracy of quarterly reports compiled by WWF and a group of more than 400 volunteers in Singapore who spend hours each month scouring local e-commerce platforms for signs of illegal wildlife trade, said Ms Jayasri, head of education and outreach at WWF Singapore.

Photos and text from posts flagged by cybercriminals are sent to an artificial intelligence model that analyzes the texts and photos for signs of illegal trade in wildlife and wildlife products.

The team must avoid mistakenly flagging replicas and must look for clues in the images, text, and prices of each listing to determine whether it is indeed for an exploited animal. Listings that have high levels of these elements may be flagged by the software as truly illegal.

“It’s harder than it looks,” Ms. Jayasri said. Spotting ivory, for example, isn’t as easy as it sounds because the Schreger lines—the unique patterns found in elephant bone that guarantee its authenticity—are hard to distinguish in often blurry photos on the Internet.

While some sellers brazenly offer products from protected species online, Ms Jayasri said, others use tactics to hide their illegal offerings.

These tactics include using emojis or code words that suggest what kind of animal is for sale, or using emojis of an animal, which “makes it harder for our team to accurately identify authentic listings.”

Replica ivory can be sold for cheap, while real ivory is more likely to sell for hundreds of dollars, Ms. Jayasri said. “But sellers can play all sorts of tricks to avoid being noticed, such as lowering prices or using different code words to avoid being suspicious. The AI ​​has to take all that into account.”

The system does not replace the work of volunteers who perform final checks of reports before they are sent.

Ms Jayasri said the team fields tens of thousands of bids each year in an effort to prevent illegal trade – the second leading cause of wildlife loss globally after deforestation and one of the most profitable areas of crime.

In Singapore, the illegal wildlife trade thrives on e-commerce platforms and in large Telegram chat groups, where buyers can purchase illegal goods such as ivory and even live and exotic animals smuggled across the border.

The Animal Concerns Research and Education Society reported that the number of banned wild animals for sale here on Telegram almost doubled between 2021 and 2023, to 660 cases. The number of sellers offering banned wild animals for sale also increased from 66 in 2021 to 387 in 2023.

The end goal, Ms. Jayasri said, is for the system to be fully automated, with the AI ​​model able to search websites for illegal listings and report on its own. But the technology has a long way to go before it reaches that level of autonomy.

The international agency plans to roll out the system to other branches soon, she added.