African jungle elephants are in serious danger of extinction: their population has decreased by 30% in the last seven years mainly due to poaching, according to the results published in 2016 by the census of Paul G. Allen's Great Elephant. But the situation could change thanks to artificial intelligence. Conservation Meetrics, a project promoted by Microsoft within its initiative AI for Earth, uses machine learning to prevent the poaching of elephants in African territory, stop the illegal trade in ivory and conserve its natural habitat.
In the rainforest of the Republic of the Congo, acoustic sensors have been installed that are responsible for collecting the sounds of animals 24 hours a day, within the initiative Elephant Listening Project. Researchers at Cornell University (New York) use advanced algorithms to distinguish the elephant whiskers – from the rest of the typical sounds of the tropical forest. Then, they use the data to estimate their population, monitor their movements, try to guarantee their safety and identify them individually.
This is one of 139 projects promoted by Microsoft in 45 countries within the program AI for Earth, whose objective is the empowerment of people and organizations to seek solutions with the help of artificial intelligence to the environmental challenges of the planet. While initially the project had a budget of two million dollars, the technology company announced last year an investment of 50 million dollars in the next 5 years.
"Today's environmental problems are created by people and the solutions to them will also be created by people. Artificial intelligence allows empowering citizens to find solutions faster, "explains Lucas Joppa, Microsoft's chief environmental officer and program leader.
AI for Earth aims to find sustainable solutions in four key areas for the health of the planet and the future of humanity: climate change, loss of biodiversity, water scarcity and agricultural performance.
It is expected that by 2050 the demand for food will exceed production by more than 70%, according to Microsoft. Farmbeat is a project that uses the machine learning algorithms of the technology company along with low cost sensors and drones to gather information about the farms. "To maximize efficiency and performance, we need to collect data on every square meter of land and thus be able to know what nutrients we should apply to a plant," explains Joppa.
Artificial intelligence to detect pathogens
"Emerging infectious diseases like Ebola or Zika are dangerous and unpredictable. Public health organizations need data as soon as possible to predict the spread of the disease and plan responses, "they argue from Microsoft. The researchers, says the company, estimate that between 60% and 75% of emerging infectious diseases come from animals.
"How do we get better knowledge about diseases in animals before they affect human society?" Joppa asks. Project Premonition is an initiative that aims to detect pathogens before they cause outbreaks. To do this, it uses mosquitoes as devices that collect animal data in the environment. There are more than 3,600 species that bite a wide range of animals, from dogs and chickens to snakes and mice: "Each bite can collect a few microliters of blood, which contain genetic information about the animal that was bitten and the pathogens that circulate in that animal "
Researchers then analyze the blood that mosquitoes take from animals to obtain valuable data: "We are developing drones that autonomously locate hot spots for mosquitoes, robotic traps to collect specimens and automatic learning algorithms to identify each bitten animal."
Among the projects promoted by Microsoft, there are some that use images of the Earth taken with drones, airplanes or satellites to guarantee the protection and conservation of ecosystems. Other initiatives seek to conserve biodiversity. "Estimates show that, without measures, 38% of all species on Earth could be extinct by the end of the century," they say from Microsoft. iNaturalist is an application supported by the California Academy of Sciences and the National Geographic Society, which offers citizens around the world the ability to collect data on different plants or animals.
When a user takes a photograph, iNaturalist uses artificial intelligence to suggest the species present in the photo. This is a useful starting point for the observer and the information collected can then be confirmed by experienced scientists. Researchers can use this data to see how populations are changing, when pest invasions occur and how species are reacting to the changing environment in real time.
The ultimate goal of AI for Earth, explains Joppa, is that any citizen anywhere in the world can through a single click to find solutions to their problems based on artificial intelligence: "We have got people to think differently about the problems and about how artificial intelligence can help solve them. "