This year marks the 50th anniversary of the Endangered Species Act. Since 1973, technology has evolved rapidly. The creation of the internet, sophisticated data management tools, gene editing, and image recognition software all came after this law passed. As these technologies continue to develop, so does their potential to help endangered species. Conservation groups already use some of these tools, but not to their full potential. Data shortages about where species live, how many there are, and the status of their habitats still present real challenges to effective species conservation.
The real challenge of technology use for conservation is implementation, not innovation. So many great ideas are missed or delayed because of timing, lack of funding and cooperation, or imbalanced resource access. There are many examples of innovative conservation groups using tools like digital crowdsourcing, machine learning, and CRISPR to help conserve species. But for these tools to truly have an impact, we will need to explore ways to help implement them at scale in order to realize impactful, positive outcomes for endangered species.
Data on Endangered Species is Neither Thorough nor Comprehensive
The International Union for Conservation of Nature (IUCN) estimates there are over 20,000 species worldwide that are data deficient—meaning there is not enough data available to determine their risk for extinction. Gathering data on species is difficult. High costs, poor technology, rough weather, and poor organization hinder our understanding of threats. Some emergent techniques, such as machine learning, have been used to try and fill this gap. Over half of all data deficient species are likely to be threatened by extinction.
The technology that can enhance data management already exists. There are many examples of small-scale uses of machine learning or imaging to help solve this problem. But they could be better implemented to help conservation experts achieve their goals. Increased use of emerging technologies would advance data collection and management far beyond where it is now.
Crowdsourcing is providing more data on conservation
We have already seen scientists use crowdsourcing tools to enhance species conservation. Crowdsourcing relies on individuals to help find and add data to programs. Rainforest Connection is one such program that began by using donated phones connected to solar panels to create cheap listening devices called guardians. They now use high sensitivity microphones to listen to the surrounding area. Computers gather the audio data from the phones, where a machine learning algorithm sifts through the data for sounds of chainsaws, vehicles, and gunshots. Having more data helps authorities identify the areas most at-risk for poaching and illegal logging. They can then focus efforts and resources in those areas and spend less time looking at data. In the Aoos Gorge of Greece, this program detected five illegal gunshots in as many months and alerted local authorities in real time. All it takes is one device for every three square kilometers of rainforest.
iNaturalist is another example of crowdsourcing helping gather data. With this app, users can take photos of wildlife, identify the species with the help of machine learning, and share their observations with a global community of over one million scientists and citizens. Over 700,000 research-grade observations since 2008 have expanded our understanding of over 10,000 data-deficient species without large increases in costs.
Machine learning helps make conservation data meaningful
Crowdsourcing can also be combined with image recognition and deep learning tools to help gather and sort more data on species. These tools can recognize aspects of images and identify wildlife much quicker than humans can, sorting through thousands of images in less than a day. SLOOP is an example of a learning algorithm that can identify a species from images collected through crowdsourcing. New Zealand’s Department of Conservation tracks a local lizard species called skinks. The use of this technology has helped improve data collection for the skinks, which gives conservationists more tools to help the species thrive.
Recent advances in imaging technology also help conservationists to improve data collection and management. Images from satellites and drones have a high enough resolution that deep learning tools can recognize species like whales and elephants from the sky. As image resolution improves, the increased use of drone and satellite cameras will greatly improve conservation efforts.
Drone imaging and machine learning can also enhance endangered species protection. In areas of Africa, WILDLABS uses artificial intelligence, drone surveillance, and previous poaching data to predict the most at-risk areas for poaching activity. Rangers then know where to focus their patrolling to remove illegal traps and equipment. This protects endangered species in areas where poaching is a major threat. This technology could be implemented in other areas across the globe to help conservation.
Increased Use of New and Emerging Tech will Assist Conservation
Besides existing tools there are also emerging technologies that may benefit endangered species. Innovations in machine learning, for instance, can be combined with other tools to foster better outcomes, or simplify data analysis that previously required human analysis to track. A self-organizing map is an unsupervised learning tool that can simplify the analysis of large, open-ended datasets and can help monitor species as the amount of data continues to grow.
Another subset of machine learning, the decision tree, is useful in classifying system outcomes as well. It does this by creating a network of two-way decisions that lead to diverse outcomes. This can help conservation by decreasing the required time for the process of listing or delisting species under the act, which currently can take more than a year to complete.
A decision tree could be part of a program that monitors current data of species and recommends listing it if threats are significant. This could streamline the review of public petitions and tracking of US Fish and Wildlife Service (FWS) processes. That would then reduce delays for species that now must wait for protections under the ESA. It could also reduce man-hours by alerting experts to the most at-risk species at a given moment. Machine learning tools, when paired with human supervision and innovation, are a game changer for efficiency of conservation data management.
CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) technology also has potential to help. It is a cheap gene editing tool that could increase species resilience or bring back extinct species like the West African black rhinoceros, which died out due to poaching. The Hawaiian honeycreeper — a small, tropical bird native to the Hawaiian Islands — has been devastated by avian malaria since it was brought to the island by explorers in the early 1900s. The disease is spread by mosquitoes. However, CRISPR makes it possible for scientists to remove the threat of nonnative mosquitos to the honeycreeper’s livelihood. The use of CRISPR-based gene drives should still be carefully evaluated. While it is possible to cause the extinction of the mosquito, for instance, a more subtle approach would seek to increase resistance to mosquito-borne parasites or diseases. It would also prevent spread to other organisms.
Other uses of CRISPR, such as SHERLOCK can quickly identify species. This could help in data collection and tracking. Currently, it is used to track the endangered delta smelt fish, which is difficult to distinguish from two other local fish without a more invasive test. With SHERLOCK, researchers can identify a delta smelt without needing days to analyze its DNA. They only need a surface swab and a 20-minute test. This process, called genetic barcoding, is where scientists add a harmless micro-mutation to the genetic structure of the species. It is then a simple process to identify it because of the unique gene sequence. This technology has potential to improve data collection and tracking of many wildlife species.
Conclusion
Innovative conservation technologies will require innovative solutions to help them become widespread. To realize the full benefit of endangered species conservation, we need more than just new ideas and technology, we need to apply them to the most pressing conservation priorities. A 2021 survey conducted by the Society for Conservation Biology on conservation groups found that the main barriers to implementation are lack of adequate funding, lack of cooperation, and insufficient capacity building. In other words, most conservation groups don’t have the resources to scale these technologies that could help endangered species.
One way to overcome these barriers is through increased partnerships between government agencies and private groups. For example, the Rainforest Connection collects data that government officials can then use to detect and prevent poaching. Another example is the Local Environmental Observer (LEO) Network, which is made up of local observers and topic experts who upload observations about unusual wildlife and environmental events. The Environmental Protection Agency relies on the LEO for valuable data that helps Arctic communities, researchers, and conservation officials alike.
When government agencies partner with private groups, together, they are able to increase the use of technology in conservation. We should continue to explore ways to implement technology like CRISPR and machine learning to help improve species conservation efforts for the next 50 years of conservation.