Better health conversations: Research on a "wayfinding" AI agent based on Gemini
4 hours ago
- #User Experience
- #Conversational AI
- #AI in Health
- Introduction of a 'Wayfinding AI' based on Gemini to improve health information accessibility through proactive conversational guidance.
- Challenges in online health information navigation: generic, overwhelming, and impersonal experiences.
- Large language models (LLMs) potential to tailor health information but current AI tools act as passive question-answerers.
- Design principles of Wayfinding AI: proactive conversational guidance, best-effort answers, and transparent reasoning.
- User study findings: participants preferred the Wayfinding AI for its helpfulness, relevance, and tailored conversations.
- Wayfinding AI's two-column interface design to separate interactive conversation from informational content.
- Results show longer, more detailed conversations with Wayfinding AI, especially for understanding symptoms.
- Conclusion: AI in health should move beyond passive question-answering to active conversational partnerships for better user experience.