WIldGuard AI

This submission, named WildGuard AI, is an AI-powered wildlife application designed to assist individuals in identifying wildlife, comprehending ecological risks, and expanding their knowledge of the natural world. The application employs a multi-agent architecture to manage various wildlife-related tasks, such as species identification, geographic verification, risk assessment, ecological knowledge, and report generation.

The primary objective is to harness AI to inspire people to explore nature, learn about wildlife, and gain a deeper understanding of ecosystems surrounding them. Unlike traditional AI applications that are confined to screen-based activities, WildGuard AI aims to serve as a valuable companion during real-world wildlife exploration.

To develop WildGuard AI, the creator utilized React, Django, Google ADK, and Gemini API. The project adheres to a multi-agent architecture, with dedicated agents responsible for specific tasks:

1. Species Identification: Identifies wildlife based on available information.

2. Geographic Verification: Determines if a species is consistent with the reported location.

3. Risk Assessment: Evaluates potential wildlife-related risks.

4. First Aid Generation: Provides relevant guidance for wildlife encounters.

5. Ecological Knowledge: Explains species characteristics and their ecological significance.

6. Report Generation: Organizes wildlife-related findings into structured reports.

The orchestration of these agents enables the completion of tasks within the application. This challenge focuses on exploring how open-source AI models can enhance wildlife identification and nature exploration, making them more accessible, flexible, and less reliant on proprietary AI services. The developer aims to emphasize the significance of open innovation, which grants developers the freedom to experiment with AI models, inspect their behavior, and tailor them to specific use cases.

In the context of wildlife applications, open-weight models provide the potential for local inference, offline identification, and greater control over how observation data is processed. Moreover, open AI models enable developers to compare various models, assess their performance on local wildlife species, and construct applications without being permanently bound to a single AI provider.

The ultimate goal is to leverage open AI for wildlife awareness and encourage people to forge a stronger connection with nature.