Poaching has been a threat for many years but the situation is now so critical that elephants and rhinos may become extinct within our lifetime. The population of Black Rhino has fallen 97.6% since 1960 and 35,000 African Elephants were killed in the last year. Despite the ongoing effort of wildlife organizations, there’s been a recent surge in poaching, which has been attributed to increased wealth in Asia. In traditional Chinese medicine the rhino horn is thought to cure impotence, fever, cancer and hangovers –there is no proof that it cures any of these ailments but this hasn’t stopped a lump of keratin becoming worth more than its weight in gold. Another motivation for the increase in poaching is a global rise in the price of Ivory, it is now worth up to $1000 a kilogram and 70% of this ivory is sold in China.

Rangers are still the most direct way of tackling the poaching problem but most patrolling teams are small and have huge areas of land to cover so don’t catch many of the traps set by poachers. They also tend to follow the same patrol routes making it easy for poachers to learn where not to hide their traps. This is where Protection Assistant for Wildlife Security (which coincidently or perhaps very deliberately – abbreviates to PAWS) comes in. PAWS is a form of artificial intelligence designed to create more efficient routes for rangers to follow, giving them a competitive advantage as they can invest their limited resources into areas where they are most likely to catch poachers. Inputting information about the protected area and previous information about patrolling and poaching activities into an algorithm produces these routes. The algorithm is based on game theory (similar to that used to create opponents in online games such as poker) and a model of poachers’ behaviour, generated by ordinary subjects and experts playing a ranger vs. poacher computer game. Data from the game was used to predict poaching hotspots and how poachers will adapt and react when different patrol routes are used, thus generating a model of their behaviour. Overtime, extra data on poaching in the area is collected allowing PAWS to learn more about the behaviour of the poachers and provide rangers with even more effective routes.

PAWS was first trialed in Queen Elizabeth National Park Uganda, the results from this test are now starting to come in and it seems PAWS is doing exactly what its supposed to – in the first month of the trial rangers found 10 antelope traps and elephant snares waiting to be released, which is far more than rangers could expect to find without the assistance of technology. However PAWS is not without faults, poor Internet and mobile signal could prevent rangers from using the software altogether. PAWS also puts rangers in danger as they are more likely to come into contact with poachers and it is not unheard-of for poachers to turn their guns to the rangers. In Virunga National Park in the Democratic Republic of Congo, poachers have killed 150 rangers in the last 10 years; in parks using PAWS this figure could be even higher.

Undeterred by these potential setbacks, Team Core, the research group that made PAWS is applying computational game theory to a range of problems. For example improving the allocation of officers and canine units in airports and transit systems to further reduce acts of terrorism. The application of readily available technology to a range of scenarios is both creative and resourceful; similar thinking would not go amiss throughout the rest of the scientific community and could generate a range of solutions to the problems our society faces.

About the author

Kate Dearling

Leave a Reply