Learning from Elders: Making an LLM-powered Chatbot for Retirement Communities more Accessible through User-centered Design
Project Overview
The document details the creation and deployment of a large language model (LLM)-powered chatbot aimed at improving accessibility and technology literacy for older adults in retirement communities. By employing a human-centered design approach, the project emphasized user feedback to develop a conversational agent that simplifies task management for its users. Utilizing GPT-3.5 Turbo, the chatbot facilitates personalized interactions that enhance user engagement with digital tools and effectively addresses prevalent challenges associated with technology use among the elderly. This initiative not only demonstrates the potential of generative AI in education but also highlights its role in bridging the digital divide, fostering greater inclusivity and empowerment for older populations in navigating modern technology.
Key Applications
LLM-powered chatbot for retirement communities
Context: Retirement community residents, particularly older adults with low technology and eHealth literacy.
Implementation: Developed using a human-centered design approach, involving user interviews and persona development. Implemented using GPT-3.5 Turbo integrated with a custom-built web interface.
Outcomes: High resident satisfaction and ease of use, improved engagement with technology, and tailored assistance for common tasks.
Challenges: Initial interface complexity, need for accessibility improvements (e.g., larger font sizes, voice-to-text feature), and concerns about data privacy.
Implementation Barriers
Technological barrier
Low technology and eHealth literacy among older adults hinder engagement with digital tools. Complex interfaces and small font sizes make it difficult for older adults to use existing technology.
Proposed Solutions: Develop user-friendly interfaces, provide tailored guidance, conduct introductory workshops to build confidence in using technology, and incorporate adjustable font sizes and high-contrast themes in the chatbot interface.
Privacy concern
Residents expressed the need for clearer privacy notices regarding data usage.
Proposed Solutions: Implement transparent privacy policies and user education on data handling.
Project Team
Luna Xingyu Li
Researcher
Ray-yuan Chung
Researcher
Feng Chen
Researcher
Wenyu Zeng
Researcher
Yein Jeon
Researcher
Oleg Zaslavsky
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Luna Xingyu Li, Ray-yuan Chung, Feng Chen, Wenyu Zeng, Yein Jeon, Oleg Zaslavsky
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai