Why the decision matters
In a real street environment, a location that appears convenient may not be the safest place to stop. Risk varies with context, and partial data should not be presented as certainty. This project asks how an AI-assisted taxi interface could communicate those uncertainties to a passenger or operator.
Prototype workflow
The interface brings together a map, nearby risk reports, uncertainty zones, explanations and alternative handoff locations. Instead of issuing only a single recommendation, it lets users explore why a location is flagged and examine nearby options that may be more suitable.
Responsible AI choices
The design emphasizes understandable explanations and visible boundaries. A user should be able to recognize when information is incomplete or a recommendation is conditional. That matters for systems that influence decisions about people and public spaces.
Implementation
The interactive prototype uses React, TypeScript and Leaflet to create an explorable map interface. The visual experience is intended for illustrating decision-support behavior and ethical trade-offs, not for controlling a vehicle or validating on-road driving safety.
Limits and possible evaluation
The prototype alone cannot establish real-world safety improvements or the accuracy of underlying risk estimates. A stronger evaluation would require validated local data, clearly defined safety criteria, and user studies that test whether explanations actually improve decisions.