Chris Seekings investigates how partnerships between NGOs and giant technology companies are harnessing the power of artificial intelligence to unmask the secrets in nature
The World Wide Fund for Nature (WWF) recently asked UK families to help scientists monitor biodiversity trends by identifying species in their gardens using iNaturalist's Seek app. Participants were able to measure everything from birds and insects to plants, flowers and fungi through the app's artificial intelligence (AI) technology, with the findings used to build a better picture of the nation's wildlife.
This is just one example of AI's use in conservation; far larger projects, involving satellite imagery and machine learning, are giving scientists unprecedented insights into the natural world. “The application of AI in wildlife protection is full of infinite potential,“ a WWF spokesperson tells me. “It can not only improve protection efficiency, but also help protection managers better understand the implementation of their own protected areas.“
While thousands of new species are discovered each year, it is thought that there are millions more waiting to be found. Much of the planet, and animals' interaction with it, remains a mystery – but NGOs are now harnessing the power of AI to shine a light onto these great unknowns.
Tech for tigers
In 2018, the WWF and tech giant Intel launched a new project to monitor wild Amur tigers in China as part of an effort to double their numbers by 2022 – the next Chinese year of the tiger. In 2010, there were thought to be just 20 living in China; around 550 exist today.
Previously, monitoring animals involved sifting through hundreds of thousands of images caught by camera traps. These camera traps are often triggered accidentally by a moving branch, leading to many blank frames. Intel's AI technology identifies whether a tiger and or another wild animal is in shot before the image is captured, ensuring every picture is useful.
“The field monitoring of wild Amur tigers is the basic premise to ensure population recovery,“ says Peiqi Liu, WWF China's head of Northeast China Project. “AI helps us identify and record the monitoring information of the tigers in a fast and real-time manner, reducing the heavy workload of manual screening and effectively improving monitoring efficiency.“
Perhaps the most impressive feature of Intel's AI technology is its ability to identify individual tigers. Each tiger has a unique stripe pattern, much like a human fingerprint. Conservationists have previously had to compare these patterns against a catalogue of images to identify them – a time-consuming and costly process – but Intel scientists have 'trained' the AI technology to recognise individual tigers from their stripes. “AI's biggest advantage is that it can achieve high accuracy recognition ability through massive self-learning, so as to liberate the time of protection workers and assist them and scientific researchers to carry out protection work more efficiently,“ Liu says.
“We hope that with the support of AI technology, wildlife protection can become more efficient“
AI for Earth
These partnerships are particularly important when you consider the International Union for Conservation of Nature's goal of assessing at least 160,000 species by 2020. Microsoft has been empowering NGOs with AI and cloud tools for years. It launched its AI for Earth initiative in 2017, and has since dished out more than 300 grants to research teams leveraging AI to monitor and manage the planet's natural systems.
“We know so little about the natural world around us because we lack data at the right level of granularity and accuracy,“ Microsoft chief environmental officer Lucas Joppa has said. “Ultimately, this means interventions are often made on good intentions and best guesses alone. The planet needs better than a guess, and researchers and practitioners need the right resources.“
AI for Earth's partnerships create large-scale habitat maps for imperilled species, combining aerial imagery, machine learning and AI to survey vast landscapes. Like the WWF-Intel partnership, this frees up time for researchers who would traditionally look through numerous images to monitor certain species.
One particular landmark for Microsoft was the creation of a chip solution that can plough through nearly 200m images in just over 10 minutes, at a cost of US$42. These results pave the way for organisations to produce new, high-resolution land cover maps on infrastructure that can scale up or down for all sorts of problems around the world.
“Algorithms need to be both fast and accurate, and there's still a lot of work and testing to do on that front,“ according to Joppa. “Nonetheless, these speedy results are a good first step in empowering people to apply AI at earth scale. And, of course, land cover mapping is just one of over 100 projects in which we have invested.“
Into the deep
Meanwhile, professional services giant Accenture has collaborated with Intel and the Sulubaaï Environmental Foundation NGO to monitor, characterise and analyse coral reef resiliency in the Philippines through its Project: CORaiL initiative.
These reefs provide habitats and shelter for approximately 25% of global marine life, and also protect coastlines from tropical storms, providing food and income for one billion people while generating US$9.6bn in tourism and recreation annually. However, the reefs are being rapidly degraded by overfishing, bottom trawling, warming temperatures and unsustainable coastal development.
“The Philippines' coral reefs have suffered due to destructive fishing methods and low marine resources management,“ a Sulubaa√Ø spokesperson tells me. “Local populations depend on fish resources, so it is absolutely necessary to preserve, restore and manage the marine resources.“
Traditional coral reef monitoring involves human divers manually capturing video footage and photos of the reef; this can be inefficient and dangerous. Engineers from Accenture, Intel and Sulubaaï have strategically placed intelligent underwater video cameras, equipped with Accenture's Applied Intelligence Video Analytics Services Platform, to detect and photograph fish as they pass by, counting and classifying marine life. The data is then sent to a surface dashboard, providing analytics and trends to researchers on the ground in real-time and thus enabling them to make data-driven decisions that will help the reef progress.
“Traditional coral reef monitoring can be disruptive to marine life as divers may inadvertently frighten fish into hiding, in addition to being dangerous and time intensive,“ Accenture Applied Intelligence managing director Ewen Plougastel explains. “Divers can also interfere with wildlife behaviour and inadvertently modify survey results. These constraints are solved with our AI and edge computing technology that samples over long periods, allowing our engineers to focus on the analysis and strategy for the project – outside of the water, and remotely.“
Since being deployed in May 2019, the solution has collected more than 71,000 images, which researchers have used to gauge reef health by analysing fish populations in real-time. “AI is empowering our engineers to achieve more and learn faster when it comes to the evolution of the marine life population within a coral reef,“ Plougastel says. “What we've developed here could be used to monitor other fragile reef ecosystems, like the Great Barrier Reef and the Mesoamerican Reef. We are looking into infrared cameras which will enable videos at night to create a complete picture of the coral ecosystem.“
“AI's biggest advantage is that it can achieve high accuracy recognition ability through massive self-learning“
Endless possibilities
The engineers have also implemented the Sulu-Reef Prosthesis, an artificial concrete reef, to provide support for unstable coral fragments under water. Fragments of living coral were planted on it and will grow and expand, providing a hybrid habitat for fish and marine life.
These are just a few of the ways that AI is being used for biodiversity protection. From infrared cameras and drones to blockchain and satellite image processing, the possibilities are endless; conservationists are also using the technology to tackle poaching and deforestation. “There are many applications for AI where we can focus on social good and improving people's lives, such as monitoring data around global warming and climate change or examining the migration of animals in parts of the world where poaching is an issue,“ Plougastel explains. “But AI should be an added contributor to how people perform their work, rather than a backstop for automation.“
Although AI may seem far removed from the natural world, it provides unprecedented opportunities for solving some of society's most vexing problems, and these partnerships show how collaboration between big tech and NGOs can have positive and sustainable environmental impacts.
“At the same time, more professional teams are needed to actively join in order to make the best use of AI,“ Liu tells me. “We hope that with the support of AI technology, wildlife protection can become more efficient. We also hope to use AI technology to bring more benefits to local community residents, achieving the goal of harmonious coexistence between humans and nature.“
Image credit: Shutterstock
