The poaching and hunting of endangered animals: it’s a problem that’s seemingly out of control. Only this week Stan Kroenke, the owner of Arsenal, launched My Outdoor TV – a channel that screens lion and elephant hunts – in the UK. And at the end of July Xanda, the six-year-old son of Cecil the Lion, was killed by a trophy hunter.
The problem is so bad that elephants and rhinos, if they’re killed at the current rates of 25,000 and 1,000 respectively per year, will be extinct in just two decades. Fortunately, however, there is hope – in the form of artificial intelligence (AI).
Neurala, a Boston-based tech start-up, will next year provide AI to a fleet of intelligent drones heading to the African savannah. They’ll be tasked with tracking animals, poachers and vehicles 24 hours a day, using infrared if necessary.
The drones are equipped with a graphics processing unit (GPU), which means they can facilitate AI processing at fast speeds. This AI is capable of a) identifying poachers who may be near animals, and b) alerting a human analyst who can call on the help of local rangers.
Neurala’s intelligent drones therefore have the potential to stop poachers – wherever they may be in a huge part of southern Africa – before they kill an endangered animal.
Max Versace, Neurala’s CEO, told us that the idea came about after he was approached by the Air Shepherd programme, part of the Lindbergh Foundation, whose aim is to eradicate illegal poaching in Africa.
“Air Shepherd’s problem was that they couldn’t find a workforce to monitor the footage their high-altitude drone was capturing. They would have needed someone to stare at a screen for eight hours just to get five seconds of poaching footage,” he explained.
“So they wanted to see if Neurala’s software could enable one operator – maybe sitting in Australia during the day rather than overnight in Africa – to remotely manage a whole fleet of drones.
“Our AI, which pre-filters all the video streams, basically says: ‘Look, human – drone number 35 has something interesting on the screen. Take a look.’ So we’re able to relieve humans from watching very boring streams and looking for a needle in a haystack.”
Neurala calls its technology the Neurala Brain – deep learning neural network software that’s lightweight, performs in real-time and doesn’t require ground intervention. Ironically, it’s modelled on animal brains, which Neurala says “do more in less space and with less power consumption that the computers we use.”
“Today there is enough power in a little chip that you can begin to insert this technology into things like teddy bears, small drones and self-driving cars,” Max added. “Our idea is to reproduce aspects of animal perception and cognition in software.”
Could such software, then, be put to task on the world’s other people-caused problems?
“I have the opposite view to those who say AI is going to destroy the world; humans are going to destroy the world, not AI,” Max said. “AI is nothing more than a very important weapon in our arsenal to make things right.
“It could, for example, be used in ocean life preservation. The Pacific is littered with billions of plastic bags and floating debris, and I don’t see humans going to pick it up. This is the type of problem – beyond the reach of humans – that autonomous robotics and good AI can help solve.”