Using artificial intelligence to combat wildlife crime

Detecting illegal fishing in Congo from space

Conkouati-Douli National Park in the Republic of Congo has a huge marine protected area that combats illegal, unreported and unregulated (IUU) fishing. The park administration relies on a program called Skylight from the Allen Institute for AI to monitor its marine protected area. Normally, it's difficult to rely on satellites to identify criminal activity on boats. How do you know if a boat is fishing illegally, sightseeing or just passing through heavy shipping traffic?

To solve this problem, Skylight analyzes satellite imagery using advanced computer vision models to detect unusual boat behavior common among illegal fishers, such as suspicious rendezvous in the area. The center can then dispatch interceptors to catch unsuspecting wildlife criminals in the act. In one case, Skylight's technology led to the arrest of illegal, unreported and unregulated shark fishermen operating at night – something neither land-based radar nor officers in boats would have otherwise been able to detect.

However, the use of AI technology poses additional risks for at-risk rangers. Many rangers in less industrialized countries, such as those evaluated during a test of the Protection Assistant for Wildlife Security (PAWS) AI system in Srepok, Cambodia, according to Harvard Business Review, are extremely knowledgeable about their area and highly motivated, but also poorly trained, equipped, and paid. If AI technology significantly increases confrontation with wildlife criminals, as it does in these examples, then there is a reasonable concern that a greater number of poorly equipped, less experienced park rangers will be harmed. A higher casualty rate could reduce rangers' willingness to use AI or deter new recruits from an already deadly profession. At the same time, increasing confrontations and thus disrupting wildlife crime has significant implications for biodiversity conservation and ecosystem services.

Diploma

Today, the use of AI in the fight against wildlife crime protects the habitat and lives of many important species. If wildlife services reject AI in this area altogether, many benefits will be lost. As this case study has shown, there are a variety of tools to combat a variety of crimes in a variety of circumstances.

Still, there are legitimate concerns about some forms of AI to combat wildlife crime. For example, concerns are that AI will be implemented as a 'quick fix' solution at the expense of people. Funds need to be allocated not only to purchase the technology itself, but also to work with and train local volunteers or police in its use. As much as AI tools can help, they are only as good as the people supporting them. To be effective, AI tools need to be complemented by well-equipped, well-trained, motivated and incorruptible rangers.

With greater awareness and likely further technological breakthroughs, the future looks a little brighter for our most endangered wildlife thanks to specific AI technology.