Artificial Intelligence Turns Poachers Into Prey

“The hunter became the hunted,” that is what artificial intelligence has currently done to poachers. As you’ve seen in the news, and all over the internet, poaching has grown into a global quandary. Poachers are killing thousands upon thousands of elephants, rhinos, and other animals and it’s clear that previous efforts have not been able to stop that.

But now, artificial intelligence has come to the rescue and believe it or not, the stubborn poachers have now turned into prey! Researchers from the University of Southern California Center tackling AI in Society are using deep learning to catch hunters in near real time –just before they strike.

How Artificial Intelligence is Turning Poachers Into Prey

Source: AIpodcast

Well, to be precise, USC has been in the front line ever since in trying to protect wildlife. Among other strategies, the analysts have been employing AI and game theory to predict where next -the hunters would strike, which in a way was not 100 percent successful.

Possibly, you must have heard of cameras that use infrared technology to monitor the hunters during nighttime pursuits. The problem with that equipment is that it also includes animal’s heat maps other than focusing on the targeted humans, meaning you could lose the track of the hunters when they come close to certain animals.

Now, the analysts have created a technology that practically differentiates humans from animals, with the best part being, the AI-based tech is able to stop the hunters from carrying out an attack just as the poachers want to act.

Teaching the Machine to Distinguish Between Human Infrared and Animal’s Infrared

Source: ARMA

For many years, the biggest hurdle has been, being able to tell between a hunter and an animal using the night surveillance binoculars. So, the team at USC computer specialist engaged their propriety labeling tool to tag or code 180,000 humans and critters in infrared videos.

To fasten things up, they decided to use a modified version of the RCNN deep learning algorithm, to train the computer to tell between infrared images of people from those of animals.

Testing the “Anti-Poachers AI Machine” in The Real World

Now, the application part came, which involved deploying the algorithm to spot hunters, recorded live by drones flying over the area. At first, in both the tests that took place in Malawi and Zimbabwe, the algorithm took approximately 10 seconds to process the images –which was not a satisfactory achievement because it takes nearly less than that amount of time to end the life of a baboon, tiger, elephant or even the most sought animal -the rhinoceros.

Fortunately, with a few tweaks, which went as far as including Microsoft’s Azure cloud platform, the Systematic POacher deTecter –“SPOT,” now takes less than 3 seconds to accurately identify poachers. Well, ideally, that’s good because the hunter will not have set his target accurately enough to strike an animal.

“The Systematic POacher deTecter as is the real name of the AI based anti-poachers machine will ease the burden of having to use drones in identifying suspicious folks in infrared imagery. In fact, it will be providing reports on hunter detection in near real time,” said Elizabeth Bondi, a lead author and Ph.D. candidate in computer science at USC.

What About the Future of Poaching?

At least, with this technology, we hope that poaching will be a thing of the past. However, It’s also important that the researchers don’t just stop there because it is possible that, as usual, the hunters will up their game.

For example, what will happen if the poachers discover a way of hiding their infrared trail –like how one hides their IP address, such that the machine is incapable of noticing them? As in, it would be better if “SPOT” or an advanced version of the machine be set with evolutionary algorithms in mind.

That would mean that the machine would be able to evolve with the poachers. As they advance their tactics, the algorithms used should be flexible enough to train itself, to catch up. In fact, today, hunters are no longer the peasants that used to be, they are people with degrees, and to fully put this menace at bay, it would be better to set up systems that will not be overtaken by time.

Nevertheless, practically, as it is now, the researchers have nailed it. This is because the artificial intelligence technology employed in SPOT promises a 98 percent success rate in providing near real-time reporting of the poachers’ whereabouts.