Google has released its AI model ‘SpeciesNet’ as open source, which identifies animal species through camera trap photos. According to foreign media such as TechCrunch, Google revealed its plan on the 3rd (local time) to significantly accelerate global wildlife conservation and biodiversity monitoring using this model.
SpeciesNet is a large AI model trained on more than 65 million photo data. It includes curated data from the Wildlife Insights user community and public repository images, capable of identifying over 2,000 animal species, vehicles, empty images, and various labels. It boasts a 99.4% accuracy in determining the presence of animals and a high accuracy of 94.5% in species-level identification.
Google applied the EfficientNet V2 M architecture to this model, combining object detection and image classifier based on MegaDetector to maximize performance. SpeciesNet can process about 2,000 photos per minute even on standard laptop environments, enabling immediate use at research sites requiring large-scale data.
In particular, Google is supporting real-time species identification and data analysis by linking this technology with the Wildlife Insights platform. This platform can process 3.6 million camera trap images per hour and classifies over 3,000 animal species with 98.6% accuracy. Built in collaboration with organizations like the WWF (World Wildlife Fund) and Conservation International, this system helps researchers and conservation groups worldwide share data and detect ecological threats in advance.
AI technology is also making significant strides in the field of poaching prevention. The PoachNet developed by the research team at Cardiff University in the UK analyzes elephant GPS data to predict poaching risk areas. Additionally, systems using AI-based cameras, audio sensors, drones to detect poaching activities or identify wildlife traps in real time have been introduced. The analytical capabilities of AI in predicting poaching hotspots and optimizing resource deployment are fundamentally transforming wildlife protection strategies.
A Google representative stated, “SpeciesNet will be a practical measure against the biodiversity crisis,” and added, “We will continue to support researchers to identify species more quickly and accurately and discover threat factors early by utilizing AI and cloud technology.”
Amid increasing endangered species due to climate crisis and habitat destruction, AI technology is emerging as an essential tool rather than just an option. SpeciesNet, newly released by Google, is a representative example capable of processing vast data and immediately deploying conservation activities, expected to have a significant impact on environmental policies worldwide in the future.