Artificial intelligence in the protection of protected areas

Artificial intelligence is increasingly being used in environmental protection. Modern digital solutions help analyze environmental data, monitor ecosystem health, and improve conservation efficiency. These technologies are increasingly being implemented in specially protected natural areas. Our article provides several examples of AI applications in Russia's protected areas.

ARTIFICIAL INTELLIGENCE IN THE PROTECTION OF THE FAR EASTERN MARINE NATURE RESERVE

A video monitoring system has been launched on Furugelm Island to protect the Far Eastern Marine Nature Reserve from poachers and to observe rare birds.

The system, using artificial intelligence, was created by the reserve's management: cameras with a 360° panoramic view; Capable of recognizing objects up to 3 km away, they transmit signals to security posts and inspectors' mobile devices.

In the spring, the cameras will be trained on the nests of the red-breasted spoonbill—in Russia, these birds breed only on Furugelm Island. Currently, there are about 30 pairs there. A screen with a live feed is planned for visitors to the island, allowing them to observe the birds' lives without disturbing them.

AI ASSISTANT FOR FAR EASTERN LEOPARD CONSERVATION

Researchers from St. Petersburg State University, together with specialists from VTB, have developed an AI assistant for analyzing archival data from Land of the Leopard National Park.

The program processes a 50-year archive of data from the Chronicle of Nature, helping analyze the state of ecosystems and select locations for camera traps. In the future, the system is planned to be expanded to include recognition of audio recordings and photographs of animals.

The Far Eastern leopard is the rarest big cat on the planet. Thanks to the creation of the national park, its population has grown from 35 individuals in the early 2000s to more than 120 in 2025.

NEURAL NETWORKS IN ANIMAL MONITORING

Neural networks also help analyze camera trap data. The systems recognize animals in photos and videos, even if the image is of low quality, allowing for faster population estimates and the identification of poachers.

These solutions are being developed by specialists from the Moscow Institute of Physics and Technology in collaboration with the Russian Ministry of Natural Resources. In 2025, the system was tested in seven Russian nature reserves; animal recognition accuracy reached 95%.

Another area is Natural sound analysis. The neural network is capable of distinguishing wolf howls from other sounds and determining pack structure.

Sber specialists collaborated with scientists from the A.N. Severtsov Institute of Ecology and Evolution of the Russian Academy of Sciences in the development. Animal voice detection accuracy reaches 98.3%, and wolf howl recognition accuracy reaches 89.6%.

The use of artificial intelligence opens up new possibilities for science and conservation—the technologies help analyze data faster and more effectively protect rare species.