Dr. Sean Yap, a research fellow at the National University of Singapore, is leveraging artificial intelligence to monitor biodiversity through bioacoustics. Working at the Lab for Advancing Protection of biodiversity with Innovative Solutions, Dr. Yap combines the study of animal sounds with AI to analyze vast audio datasets collected from tropical forests. This approach allows researchers to identify species, track ecological patterns, and locate individual animals without relying solely on visual observation. The primary objective of this research is to understand how human-generated noise, such as traffic, impacts animal activity and to evaluate whether small restored forest patches, known as microforests, can enhance ecological connectivity in urban landscapes. To gather data, the team deploys autonomous recorders equipped with specialized microphones that capture sound frequencies across the forest environment. These recordings are converted into spectrograms, visual representations of sound that AI systems analyze to distinguish between broad categories like traffic noise and specific animal calls. Unlike traditional field surveys, which depend on human observers and are limited by duration and subjectivity, bioacoustic monitoring offers continuous, standardized data collection. Recorders can operate 24 hours a day, capturing species that might be elusive or avoidant of human presence. This method provides a more comprehensive view of animal activity patterns and disturbance levels over extended periods. However, the technology faces specific challenges. AI models currently perform best with species that have distinctive vocalizations, such as songbirds, but struggle with lower-frequency sounds produced by pigeons, doves, and owls. In some instances, the system has misidentified traffic noise as animal calls. A significant limitation is that many existing sound-recognition algorithms are trained on data from North America and Europe, reducing their accuracy for the unique biodiversity of Southeast Asia. To address these gaps, researchers at NUS are refining models using locally collected data. As regional biodiversity datasets expand, the AI systems are expected to become more reliable. Dr. Yap emphasizes that AI serves as a powerful complement to traditional research rather than a replacement. The technology relies on high-quality training data and the ecological expertise of scientists to interpret findings accurately. In biodiverse regions like Southeast Asia, where dense forests make extensive field surveys difficult, bioacoustics enables the collection of larger, more consistent datasets. This capability is crucial for understanding ecosystem function and designing effective conservation strategies. Looking ahead, Dr. Yap aims to expand AI-based monitoring beyond birds to include frogs, insects, bats, and other mammals. By developing locally trained algorithms, the team hopes to improve the accuracy of biodiversity monitoring and support conservation efforts in the region.