The illegal trafficking of marine wildlife moves billions of dollars each year and represents a direct threat to endangered species. Among the most trafficked items are shark fins, seahorses, and sea cucumbers, which are often hidden in luggage or packages and cross borders undetected.
Dr. Vanessa Pirotta from Macquarie University (Australia), the lead author of the study published in Frontiers in Ocean Sustainability, stated: “The wildlife trade is cruel and unethical. We take this World Oceans Day to highlight this issue.”
Artificial intelligence as a detection tool
To address this issue, scientists at Macquarie University developed an artificial intelligence algorithm capable of detecting samples of these species with an accuracy of 92%.
The system repurposes X-ray computed tomography (CT) scanners already present in airports, which generate 3D images of the contents of suitcases. Through a neural network trained with hundreds of scans, the algorithm can automatically identify suspicious items and flag luggage for inspection.
How the algorithm was trained
Researchers conducted 298 scans of 20 sea cucumber samples, 30 seahorse samples, and 18 shark fin samples, many from actual seizures.
- Smuggling tactics were simulated, such as wrapping samples in clothing, cans, or toys.
- Images were added to scanned suitcases without illegal goods using the Threat Image Projection technique.
- The algorithm was tested with a subset of images never seen before.
The results were conclusive:
- 95% accuracy in detecting shark fins.
- 96% in seahorses.
- 86% in sea cucumbers. The false positive rate was 13%, with variations depending on the species.
Limitations and challenges
Although the accuracy is high, the system is not a definitive solution:
- False positives require manual checks.
- Not all airports have 3D CT scanners, as they are expensive; many still rely on 2D scanners.
- The algorithm must be complemented with human detection and the use of sniffer dogs.
“AI is not the definitive solution for detection, nor a substitute for human detection,” concluded Pirotta.
Global implications
The application of artificial intelligence in the fight against marine wildlife trafficking opens up new possibilities for:
- Interception of illegal shipments.
- Cutting commercial routes that supply human, medicinal, and ornamental consumption markets.
- Protection of vulnerable species, reducing pressure on precariously balanced populations.
- Prevention of invasive species, preventing trafficked live animals from escaping and altering ecosystems.
The development of this algorithm represents a significant advancement in the protection of marine biodiversity. Although it does not replace traditional methods, it does offer a powerful tool to reinforce controls at airports and borders.
In a context where illegal trafficking threatens the survival of entire species, the combination of technology, international cooperation, and public awareness will be key to addressing this global challenge.
