In a recent report, the UK's Nature website pointed out that an increasing number of scientists are using artificial intelligence in the field of biodiversity conservation. They analyze a large amount of data, monitor ecosystems, identify trends over time, address the challenge of biodiversity loss, and save endangered species. Artificial intelligence bears the heavy responsibility. Currently, the rate of species disappearance is tens of thousands of times higher than millions of years ago, with up to one million species on the brink of extinction. In view of this, the United Nations set a goal in 2020: by 2030, at least 30% of the Earth's land, freshwater, and oceans will be protected or preserved in some form. So far, only about 15% of the world's land and 7% of the ocean have been protected to some extent. Nicholas Meay, the founder of the international non-profit organization "Future Society," stated that although artificial intelligence is currently "not perfect," it may accelerate important discoveries and require the participation of artificial intelligence experts in designing models, as well as collecting, labeling, quality checking, and interpreting data. Conservation AI's tools can identify species from the camera lens (Image source: Nature website). The UK non-profit organization Conservation AI is using artificial intelligence technology for various ecological projects. Experts from the organization say that without artificial intelligence, humans may never be able to achieve the United Nations' goal of protecting endangered species. The study by Joeg Miller, an ecologist at the University of Werzburg in Germany, has demonstrated that artificial intelligence tools can identify animal species from recordings, thereby helping to quantify the biodiversity of tropical forests. In a study published in the journal Nature Communications on October 17th, scientists used artificial intelligence to analyze the "soundscape" of animals in the Choc ó region of Ecuador, known for its rich species diversity. They placed recorders on 43 pieces of land representing different stages of restoration, recording the 'soundscape' of animals. These lands include forest land that has not been affected by deforestation, areas that have been deforested but subsequently abandoned and begin to regenerate, and land used for cocoa plantations and pastures. Subsequently, they handed over the recorded audio files to experts, who identified 183 species of birds, 41 species of amphibians, and 3 species of mammals. The researchers also input their recordings into a Convolutional Neural Network (CNN), which can recognize the sounds of birds. The results showed that CNN was able to identify 75 out of the 183 bird species identified by experts, but the dataset of this model is limited and only includes 77 possible bird species that may occur in the region. Artificial neural networks can identify birds from rainforest recordings (Image source: Nature website). This study shows that artificial intelligence can use sound to more comprehensively identify species in tropical regions, and what is needed now is for humans to collect more data. Researchers from the Conservation AI organization for real-time monitoring of biodiversity have developed models that can search for shots and images captured by drones or infrared cameras to identify wild animals (including critically endangered species) and track their activities.