Google Expands Wildlife Monitoring With Open-Source SpeciesNet

Google published "How our open-source AI model SpeciesNet is helping to promote wildlife conservation" on March 6, 2026, presenting real-world deployments of its wildlife-identification model across Africa, South America, North America, and Australia.

SpeciesNet is designed to classify animals in motion-triggered camera-trap photos, a workflow that normally requires enormous manual effort. Google states that the model can identify nearly 2,500 categories of mammals, birds, and reptiles. The company says SpeciesNet has been used through Wildlife Insights since 2019, and that the model was released as a free open-source tool about a year before this update.

The post provides concrete field examples. In Tanzania, the Snapshot Serengeti project used SpeciesNet to process a backlog of 11 million photos in days, turning what had been a major labeling bottleneck into tractable research throughput. In Colombia, the Humboldt Institute and the Red Otus network used camera-trap analysis to track behavioral shifts, including changes in bird migration timing and indications that some mammals are becoming more nocturnal under pressure. In Idaho, the Department of Fish and Game uses SpeciesNet to pre-sort millions of images each year so human experts can focus on final review. In Australia, WildObs trained localized variants to cover species that were not in the original model, including region-specific conservation priorities.

These examples matter because they show where open models can produce immediate public-value gains: faster ecological monitoring, broader geographic coverage, and improved continuity of long-term biodiversity datasets. For conservation agencies, model adaptation and human-in-the-loop validation remain essential, but automation of first-pass classification can dramatically change operational scale.

Google positions SpeciesNet as infrastructure for practical biodiversity intelligence rather than a one-off demo. If adoption continues, the largest impact may come from standardizing how wildlife evidence is collected, labeled, and shared across institutions that previously lacked compatible tooling.