Inspiration
Women safety during ride-sharing is often reactive, relying on SOS triggers. I wanted to build a system where AI proactively reasons about ride risks before emergencies occur.
What it does
AI Safety Incident Commander is an autonomous AI agent powered by Gemini 3. It continuously monitors live ride telemetry such as route deviations, delays, speed anomalies, and unexpected stops to detect emerging risks.
How we built it
The system uses Node.js and Express to send structured ride data to Gemini 3 via Google AI Studio. Gemini acts as a long-running Incident Commander, maintaining ride context and autonomously deciding safety actions using machine-readable JSON outputs.
Challenges we ran into
The main challenge was designing Gemini as an orchestrator rather than a chatbot, ensuring continuous reasoning and autonomous decision-making instead of single-prompt responses.
What we learned
This project represents the Action Era of AI, where models are not passive responders but active agents that monitor, reason, and act to protect users in real time.
Built With
- 3
- ai
- apis
- autonomous
- express.js
- gemini
- javascript
- json-based
- node.js
- rest
- studio)
- telemetry