End of 2019, the German Federal Environment Ministry launched its new funding initiative “AI Lighthouses for the Environment, Climate, Nature and Resources”. With a total of 27 million euros, projects are to be funded here that use artificial intelligence (AI) to tackle ecological challenges. In fact, there is already a whole range of approaches to using AI to improve air and water quality, stop species loss, track down poachers or grow crops more sustainably and conserve resources.
AI in species conservation
Around 3,200 tigers currently live in the wild. The governments of all countries with tiger populations agreed in 2010 to double their numbers by 2022. In order to achieve this goal and protect the big cats in the long term, it is necessary to know where how many tigers live and the ways in which they spread. In China, the nature conservation organization WWF uses artificial intelligence to evaluate thousands and thousands of photographs taken every day by automatic photo traps. Whereas it used to be humans who had to look at and sort out the countless images, tiger photos are now recognized directly by optimized image recognition. Thanks to cloud technology, they are also sent more quickly from mostly impassable groves to experts who can evaluate them promptly. This saves time and resources. In a further step, artificial intelligence will help identify individual tigers by their personal stripe pattern. A method that U.S. scientists at the University of Illinois at Chicago are already using in the Wildbook project to determine the identity of zebras, humpback whales or whale sharks. For their part, German researchers want to develop a “weather station for biodiversity” with AI support to better protect insects and birds.
AI monitors the environment
As the global hunger for salmon fillets continues to grow, so does the number of salmon farms in coastal waters. Strict regulations and monitoring programs are designed to help ensure that the ecosystems there are not stressed beyond their limits. To date, worms, mussels, starfish, crabs and other macroinvertebrates living on the seafloor have been examined under the microscope. But this is very expensive and time-consuming and only allows the examination of a few samples. Researchers at TU Kaiserslautern are now working with colleagues from Scotland and Switzerland on a digital alternative: They want to use AI to study microorganisms in the future. These react very quickly and sensitively to a change in their environment and are therefore ideally suited as bioindicators. However, the identification of microbes is very difficult with conventional methods. Machine learning should now make it possible to exploit the potential of microbes as bioindicators. This will allow faster, cheaper and more frequent recording of how a salmon farm is doing in terms of ecosystem health. But until that happens, the algorithms will need more training. That’s because the composition of bacterial communities in the seafloor can vary greatly depending on the time of year, location and local conditions. Current research is therefore focusing on integrating these influencing variables into the machine learning algorithm. In this way, it will be gradually refined until it is finally able to monitor environmental pollution in an automated way – to support the careful use of our natural resources.
AI could reduce food waste
Eleven million tons of food become waste each year in Germany already during production. Strict requirements for product safety, low plannability in agriculture, countless product-specific constraints in food processing, strong fluctuations in demand and the trend toward individualized products also in the food industry have so far prevented the reduction of this ecological and economic waste, explain the cooperation partners of the REIF project. REIF stands for “Resource-efficient, Economic and Intelligent Foodchain”. Various universities, research institutions and food manufacturers have joined forces to tackle the problem with AI. Funded by the German Federal Ministry of Economics and Technology, the research project will spend the next three years “investigating the potential of artificial intelligence to optimize the ability to plan and control value creation in the food industry.” The goal: to build an AI ecosystem that digitally brings all participants in food production on board, thereby reducing food waste in the future.
AI saves energy
To meet climate targets, we not only need to produce our electricity more sustainably, but also reduce our energy consumption. Artificial intelligence could also help us do this. The non-profit Borderstep Institute in Berlin, for example, is using machine-learning algorithms to regulate heating in a Berlin neighborhood as part of its WindNODE showcase project. With the help of sensors in the apartments and buildings, the system can determine when residents are at home and turn up the heating accordingly. In the process, the system is expected to adapt more and more to the habits of the residents, thereby saving 20 to 25 percent of energy. In addition, scientists at Landshut University of Applied Sciences are working in the DENU research project to significantly reduce energy requirements through intelligent networking. To this end, the researchers are installing measuring and control devices in hotels, indoor swimming pools, factories and office buildings in Lower Bavaria, for example, and linking the measured data with existing energy efficiency management systems to create a holistic system. Machine learning is then used to analyze the collected data and develop algorithms to reduce the buildings’ energy consumption through intelligent control. “By looking at all factors holistically, we can save more than 50 percent of primary energy,” explains professor and project manager Diana Hehenberger-Risse. The project, which is funded by the German Federal Ministry of Economics and Technology to the tune of 1.4 million euros, aims to determine by 2022 whether this estimate is actually correct and whether the model developed in Landshut can become a blueprint for the intelligent and resource-saving energy management of the energy turnaround.
AI to make agriculture more efficient and environmentally friendly
It is predicted that by 2050, humanity will need around 70 percent more food than we produce today. Better and more accurate data could help farmers grow food more efficiently while protecting the environment. Microsoft’s FarmBeats project, for example, feeds data from sensors, drones, satellites and tractors into cloud-based artificial intelligence models to provide a detailed picture of soil quality and field moisture. Since fast internet is a rarity on most farms, the data is sent to the cloud via unused broadcast frequencies between TV channels. Together with the United States Department of Agriculture, FarmBeats has now launched a pilot project: On the ministry’s approximately 2,800-acre research farm in the U.S. state of Maryland, fields were equipped with a network of sensors for this purpose. They measure the temperature, moisture, acidity and water level of the soil. A weather station will record air temperature, precipitation and wind speed, and a tractor equipped with sensors will in turn measure the height, biomass and “greenness” of the crops – an indicator of plant health. If the project is successful, farmers will be able to see the data generated by FarmBeats in real time. Researchers, on the other hand, will be able to provide farmers with web-based tools and site-specific information that will allow them to better target seeds and fertilizers and refine their farming practices overall. If all goes according to plan, the system will subsequently be tested more extensively on more than 200 farms across the country.