New research shows that AI-generated images can help improve conservation efforts by supplementing real-world data, but they cannot fully replace actual observations. Scientists at North Carolina State University tested whether synthetic images of 20 common North American trees could enhance the performance of image-recognition models, especially when real images are scarce. They combined real images from sources like iNaturalist and the Auto Arborist Dataset with AI-generated images and tested various combinations on a standard image-recognition model. The results showed that synthetic images helped the model recognize trees better when real images were limited, but they were still less effective than real images overall. This study was published in the journal Remote Sensing in Ecology and Conservation. Thomas Lake, the lead author of the study and a research scholar at NC State's Center for Geospatial Analytics, said the biggest benefit of AI-generated images could be in areas where data is limited. However, he emphasized that these images shouldn't be seen as a complete replacement for real-world observations. The study found that synthetic images lacked the subtle details and natural variations found in real images, which are vital for accurate species identification. Co-author Chris Jones, a senior research scholar at the same center, explained that some species are photographed often, while others are rare, live in remote areas, or are overlooked. Even for common species, photographs may be limited to specific places, seasons, or angles. AI-generated images can help expand small real-world image collections, allowing scientists to build initial models for species identification and monitoring in areas with limited observations. These models can then be improved as more real images are collected over time. Lake stressed the importance of community science in data-driven conservation efforts. Real images contributed by the public provide essential details like leaf shape, bark texture, and growth patterns, which are crucial for identifying species. He noted that while AI-generated images don’t eliminate the need for field observations or community science, they can offer new ways to use those observations for recognizing species and tracking environmental changes. Lake emphasized that real people and real observations remain central to conservation. Community science platforms like iNaturalist are powerful because they allow thousands of people to collectively observe nature in ways no single research team could. Anyone can participate by contributing photos of trees, birds, insects, or flowers, helping to document where species live and how the environment is changing. While technology can help analyze and interpret data on a larger scale, understanding a changing natural world still requires people to pay close attention and document their observations.