The World Economic Forum’s (WEF) annual January bash is known for hosting world leaders discussing the most pressing political, social and economic issues of the time. But the two stars of Davos 2019 were not the usual familiar names from the political elite like Trump or Putin – instead, it was the presence of Sir David Attenborough and Prince William in conversation about our impact on the environment that proved a showstopper moment.
And rightly so. To mark the event, WEF launched a report into electronics production, revealing that the amount of e-waste generated every year - 44.7 million tonnes - weighs more than 125,000 Boeing 747 jumbo jets in total, enough to build 4,500 replicas of the Eiffel Tower.
As well as contributing more unwanted goods to already overflowing landfill sites, the entire lifecycle of our smartphones, tablets and computers adds pressure on the planet. From industrial mining for minerals like cobalt and tantalum, which requires digging through 30kg of rock for every 100g of mineral and leaves behind a scarred landscape and toxic waste; to the coal-powered plants where the devices are made; to the rapid rate at which they’re manufactured, used and thrown away – the technology sector has a lot to answer for when it comes to detrimental impact on our planet and resources.
But enough of the bad news. Fortunately, there are also many examples where technology is giving back, with various projects aimed at protecting the environment through preventing poaching and stopping the disappearance of more endangered species.
Poaching the poachers
Poaching is one of the biggest threats to wildlife today. Elephant and rhino are both prime targets for armed poaching gangs, intent on hacking off tusks and horns to sell on the black market to be used for traditional medicines and trinkets. As the market for these items is worth millions, poaching has now become a highly professional endeavor, with individuals armed with assault weapons being flown in across borders to carry out attacks.
Luckily, technology is beginning to foil these efforts.
Back in 2015, Dimension Data teamed up with Cisco to work on ways to tackle the poaching problem, after being approached by a private reserve in South Africa – Dimension Data’s home country – which was losing a rhino a week to poachers.
The standard approach to protecting rhino is putting a collar on the animal or putting a tracker in its horn. Both of these methods interfere with the animal as they involve darting and sedation, causing stress and potentially endangering its life – many forest elephants have died after being darted, as if they slip and fall awkwardly when they run off into the bush, it crushes their internal organs.
So Dimension Data and Cisco decided to shift their focus to tracking the humans rather than the animals. They set up a network of sensors on the fences around the park, and placed seismic sensors on the ground, which would set off an alert if the fence was cut or someone came in underneath it.
They deployed thermal imaging cameras at certain points along the fence to view how many people were present, their location and the direction they were heading. Number-plate and facial recognition technology was also installed at the gates to prevent known poachers gaining access through the front door.
All the data gathered is sent to the rangers’ mobile device or iPad via a new radio network, which can reach rangers wherever they are across the entire park. Phase two of this project has seen the time it takes for a ranger to receive a report drop down from 30 to seven minutes, meaning they have a much better chance of getting to the poachers before any animals are killed.
The project has been a huge success so far. In 2015, 54 rhinos were killed by poachers; in 2017, not one rhino lost its life. The firms are now rolling out similar projects in parks in three different countries - Kenya, Zambia and Mozambique – aimed at protecting rhino and elephant.
As a biologist by academic training and a passionate conservationist, Ruth Rowan, Dimension Data CMO, is optimistic about the role of the technology sector in protecting our species and environment:
The world is very passionate about conservation. It feels like there's a very strong trend at the moment, and certainly in the tech industry but also across the board. There is an awful lot of power in large foundations and corporates to help support and empower great ideas.
Sometimes it's marrying up the execution ability that you might see – someone who has done that mapping and what they need a little bit of help with is a thermal imaging camera or the funding for that – with the right foundational partner to help support that. We have had a lot of interest from a lot of different foundations to understand how our technology expertise can help them with their project.
For their first project in South Africa, Cisco and Dimension Data donated some of the kit and future rollouts could be funded on a charitable basis, depending on the circumstances and the park. Working on technology projects to protect wildlife or the planet isn’t something Dimension Data views as a commercial opportunity or revenue generator, according to Rowan:
Many of us in the company are passionate conservationists. We also just recognize a real problem in the world, that certain animals are very much at risk of becoming extinct and we believe that technology can help solve that problem and protect animals from poachers. That's really been our motivation in getting involved.
“There is a lot of debate at the moment that technology in some instances is potentially creating more risk - the misuse of data or speculation around the manipulation of the data in the social media space to influence people – but we believe that technology is here to make the world a better place in the broadest sense. It’s just got to be used responsibly.
Artificial intelligence (AI) is another technology being directed to protecting wildlife. In January, non-profit organisation Resolve launched TrailGuard AI. The security camera system can be hidden from poachers due to its small size – similar to a typical index finger – has 1.5 years of battery life, and features human-detection algorithms run directly from an Intel-based processor housed in the camera head.
The AI aspect of the camera means that rangers are only alerted when motion is detected that has a risk of poaching. As an elephant is killed every 15 minutes by a poacher, at a rate of approximately 35,000 elephants per year, this more accurate alerting system is vital to ensuring rangers are directing their efforts to the right places.
WildTech@Resolve plans to provide TrailGuard camera systems to national parks, conservation groups and scientists in 100 reserves across Africa throughout this year, with plans to expand to Southeast Asia and South America, aided by support from Intel, the National Geographic Society and the Leonardo DiCaprio Foundation.
In January, the British International Education Association (BIEA) hosted its first STEM conference, aimed at exploring how AI and other technology can help protect vulnerable species and ecosystems.
At the event, Dr Josh Veitch-Michaelis, postdoctoral research assistant in the Astro-Ecology Group at the Astrophysics Research Institute, Liverpool John Moores University, discussed his team’s project to build a system capable of identifying animals in real-time on a drone, as a means of data-gathering for wildlife conservation.
With less than half as many wild animals now as there were 30 years ago, Veitch-Michaelis explained the vital need to locate animals and do accurate animal counts in certain areas, to investigate whether numbers are dropping due to habitat destruction, poaching or other factors.
However, current animal-counting methods based on field trips or even using cameras attached to drones aren’t adequate to do this accurately. The footage captured by these cameras still requires manual intervention, with someone sat at a computer somewhere behind the scenes receiving the images, deciding whether there are any animals present, and then counting and recording the results. Veitch-Michaelis explained:
This is quite labor-intensive. What we really want is something where we can take an image, put it onto a computer, and have the computer label the image for us, which would save us going through thousands of images ourselves.
Using normal cameras can also lead to false detections – a rock might look like a rhino - and can lead to over-estimating the population.
Veitch-Michaelis is part of an Astro-Ecology Group made up of ecologists, astrophysicists, computer scientists, engineers and drone pilots, who have teamed up to combine computing with the same techniques used in astronomy to solve these problems.
The team are using FLIR thermal imaging cameras attached to drones that process visual data via a heat map, making it much easier to differentiate animals who radiate heat from a stone cold rock, and the results are much more reliable. This system can also be used for spotting poachers.
But even with this advanced camera technology, there is still potential for false detections. Veitch-Michaelis said:
Animal waste is the same temperature as the animal. When you’re looking at a drone-based video it’s quite easy to see these bright spots and think that could be a human, and it’s not. We need ways of identifying these false detections.
This is where AI comes in. To accurately label an image showing hundreds of heat dots would normally suck up a lot of student time, or paying people to carry out the task for millions of images. The Astro-Ecology Group is instead using a combination of machine learning and citizen science, based on a custom-built AI system, explained Veitch-Michaelis:
We’re experimenting with Raspberry Pis to capture and process data along with Nvidia GPUs for training our models. We use some common frameworks like the Robot Operating System, OpenCV and Pytorch/Tensorflow, but the software on top is all custom.
They are training the system to identify and label animals in images, rewarding correct classifications, penalising incorrect or missing classifications, and then modifying the model for each batch of images processed based on these results.
This work is supported by people-powered research site Zooniverse, where its one million-plus users are being encouraged to scour the same heat-map images the computer has looked at, to make sure all animals are detected, none have been missed, and that the algorithms can identify different species correctly. Veitch-Michaelis added:
This is what we think the future of ecology is. We want to build these systems to rapidly classify thousands of pictures, and ideally we want this running on a drone itself. Our goal is to build a system that can do animal identification live on a drone. We can’t assume we have enough bandwidth to stream live, high-resolution images back to the ground. We need to do the processing in the air - this is particularly important for applications like anti-poaching where you can’t wait to process the data later.
The system we want doesn’t exist and we’re trying to build something that is reasonably priced and could be replicated by other researchers. We don’t have a definitive system yet.
The BIEA is hoping to inspire a whole new generation of technologists and scientists to turn their attention to building technology systems like the one on Veitch-Michaelis’ wish list with its annual STEM competition.
The theme of this year’s International STEM Youth Innovation Competition, run in conjunction with the Born Free Foundation, is fighting extinction through technology. The aim is to show students how STEM can make a difference in solving global problems like wildlife conservation, and could have a knock-on effect of tackling the ever-growing IT skills gap by inspiring a new batch of technology experts.
The technology industry takes a lot of criticism for the world’s ills – and rightly so in many cases - so it’s good to focus on the positive ways tech is being used to try and protect our planet, especially if these projects inspire more young people to consider technology as a career.