Tackling Online Wildlife Crime with Artificial Intelligence
As online traffickers use social media with coded listings to trade endangered species, we look at how AI tools track suspicious posts, decode hidden terms, and help law enforcement seize items.
The Digital Face of Wildlife Trafficking
Over the past few years, smugglers have realised there’s a massive advantage to doing business online. Social media platforms, instant messaging apps, and e-commerce sites offer global reach, less risk of detection, and a plethora of ways to disguise illegal dealings. Instead of physically hawking ivory tusks or rare reptiles, criminals can now hide in the relative anonymity of the digital world.
They’ve become cunning too, adopting everyday code words to mask illicit transactions from unsuspecting moderators. For instance, endangered turtles might be advertised under the euphemism “freshwater fish,” while illegal ivory sometimes appears in conversation as “red flowers.” Disturbingly innocuous-sounding, these cryptic labels exist to throw off authorities – and their algorithms.
That’s where artificial intelligence steps in.
Meeting the Challenge with AI Guardian and Cyber-Surveillance
IFAW’s AI Guardian and WWF’s cyber-surveillance system—two pioneering platforms that scan online marketplaces, classify posts, and look for suspicious patterns of language that might signal wildlife trafficking. They tirelessly combe through millions of posts, adverts, and messages in search of an elusive clue: a strange phrase, a particular pattern of emojis, or a giveaway combination of words.
- Language Detection
AI tools rely on vast linguistic datasets and advanced natural language processing to weed out the normal items from the deeply illegal. They learn typical patterns like describing an animal’s size, colour, or origin. When the AI encounters “red flowers” in a description that has no business mentioning flowers—especially if that same advert has “for sale” or “genuine collector’s piece” in it—it raises a red flag. - Image Recognition
In some cases, wildlife traders try to sneak images of actual animals or body parts (like tiger teeth or rhino horn) into their sales listings. Increasingly, these images are disguised or partially hidden. AI-driven image recognition can detect these suspicious images—even when they’re cropped, blurred, or partially masked. - Pattern Analysis
Platforms like WWF’s system also track the behaviour of users: how frequently they post, which categories they sell under, and what times of day or night they’re most active. This helps create ‘user profiles’ that can be flagged for further investigation if they display the telltale signs of illegal activity.
Machine Learning: Learning to Adapt
Once a trafficker realises their code word (like “red flowers”) has been discovered, they’ll shift tactics. Instead of “red flowers,” they might switch to “antique blooms.” And that’s the crux of this ongoing war: each time criminals adapt, the AI has to adapt faster.
Machine learning underpins this capability. Every bit of new data—be it a new term, a new method of concealment, or a subtle shift in how these traffickers operate—feeds back into the AI, refining its processes and digital detection tools. The more data they’re fed, the better they become at sifting out legitimate listings from criminal fronts.
But it’s not just about language. Machine learning can pick up on the anomalies in user behaviour, for example:
- Sudden spikes in posting activity by a previously dormant user.
- Consistent misspellings or coded references to specific species.
- Geographic discrepancies (someone posting items from multiple countries within the same week).
These subtle data points, when combined, paint a picture that’s often enough for enforcement agencies to launch an investigation or even an undercover purchase to confirm suspicions.
A Success Story: Hawksbill Turtle Sale on eBay
In April 2017, automated tools monitoring eBay France flagged a listing described somewhat vaguely as a “rare maritime decoration.” On the surface, the wording didn’t directly indicate anything illegal—just another seaside-themed item that might catch the eye of a collector. However, images attached to the post revealed a distinctive shell pattern. Although not mentioned explicitly in the text, the shell in the photos appeared to match the marbled, overlapping scutes of a hawksbill turtle (Eretmochelys imbricata), a species strictly protected under CITES Appendix I.
Because the descriptive text didn’t align with the visuals, eBay’s filters and human moderators took a closer look. They compared the shell markings to a protected-species database and discovered a clear match with a hawksbill turtle shell. At that point, they removed the listing and alerted French wildlife authorities. Under French and EU law, buying or selling hawksbill turtles—or items derived from them—is prohibited.
The seller was located and found to have offered other similar items, though details on those transactions weren’t made public. Once enforcement action was taken, the hawksbill shell was seized, and legal proceedings began.
Why AI is a Game-Changer
This real-life interception illuminates how crucial speed and data analysis are in stopping wildlife crime. A handful of human moderators, no matter how dedicated, can’t hope to keep up with the millions of online posts generated every minute. AI, however, can be scaled rapidly to handle vast volumes of data—ensuring that suspicious activity doesn’t go unnoticed.
Moreover, AI doesn’t just function as a static filter. Its strength lies in continuous learning. Each successful detection teaches the algorithms something new about trafficker behaviour, making future detection even more precise.
A Race Against Time
Digital marketplaces may offer a haven for illegal wildlife traders today, but they also create a digital fingerprint that AI can track and trace. With AI scanning millions of online listings and messages—analysing text, images, and user behaviour in near real time—these systems can quickly adapt to traffickers’ changing tactics.
AI-led monitoring of e-commerce and social platforms is revealing coded adverts and linking sellers to known networks. By comparing suspicious posts with protected-species databases and automatically flagging high-risk users, AI provides crucial leads for law enforcement. This integrated, technology-driven approach helps dismantle digital trafficking routes, complementing traditional anti-poaching efforts and offering a stronger chance to protect endangered species.
At Cow-Shed, we believe in championing the power of technology to protect our natural world. We’ve been inspired by the determination of organisations like IFAW and WWF in pioneering AI solutions to combat online wildlife crime. As the threats evolve, so must our response. Let’s continue to support and spread the word on the game-changing tools that can help ensure these magnificent creatures don’t just survive, but thrive.
