Google has launched a cutting-edge AI tool focused on enhancing wildlife monitoring by recognizing different animal species.

On Monday, the tech powerhouse revealed **SpeciesNet**, an open-source model tailored for wildlife biologists. This tool comprises two models: one that identifies objects in footage taken by wildlife monitoring cameras and another that categorizes those objects into distinct animal species.

Since 2019, SpeciesNet has been accessible through **Wildlife Insights**, a Google Cloud-based platform for biologists. However, as of this week, Google has made SpeciesNet available to the public as an **open-source model on GitHub**.

Wildlife researchers frequently depend on motion-activated cameras to watch animals in their natural surroundings. Nevertheless, processing the enormous volume of footage collected can be a lengthy endeavor. AI can significantly expedite this process, enabling conservationists to dedicate more time to protecting wildlife instead of sifting through images. “AI can speed up that processing, allowing conservation practitioners to spend more time on conservation and less time reviewing images,” mentions the [SpeciesNet repository](https://github.com/google/cameratrapai?tab=readme-ov-file#overview) on GitHub.

As per Google, SpeciesNet was trained on a dataset of over **65 million images**, encompassing camera trap images from Wildlife Insights users, as well as publicly accessible datasets. The model evaluates data from its foundational systems to predict each detected animal, providing a confidence percentage for its classification.

SpeciesNet can categorize images into more than **2,000 categories**, spanning a diverse array of animal species, broader taxonomic groups (like “mammalia” or “felidae”), and even non-animal items such as “vehicles” or “blank” images.

The **open-source model** is now accessible for public utilization on [GitHub](https://github.com/google/cameratrapai).