AI Dementia Aid Assistant

Overview

This project was driven by the idea that technology can play a meaningful role in supporting people with memory loss. I designed a system that helps dementia patients recall important memories, stay oriented, and feel less alone through natural, personalized conversations.

Problem

Dementia doesn’t just affect memory — it impacts a person’s sense of identity, independence, and connection to others. Many existing tools are overly clinical or generic, and don’t reflect the emotional and personal nature of memory care.

Process

I designed a system that combines structured personal data with semantic memory retrieval, allowing the assistant to surface meaningful memories during conversations. A key focus was making interactions feel familiar and comforting, including exploring voice cloning techniques to model speech patterns and tone based on recordings of loved ones. Alongside the technical work, I researched communication approaches used in dementia care, such as gentle reinforcement, avoiding contradiction, and maintaining emotional consistency, and used these principles to guide how the AI responds.

Outcome

The result was a prototype that could hold contextual conversations, recall relevant personal details, and respond in a way that feels more familiar and supportive. Rather than simply answering questions, the system was designed to provide reassurance, continuity, and a sense of connection.

Lessons Learned

This project taught me that building AI for real people — especially vulnerable populations — requires more than technical accuracy. It requires empathy, careful design decisions, and an understanding of how small details in tone and response can meaningfully impact someone’s experience.