'It's like having a friend': Exploring the Opportunity of Large Language Models (LLMs) in Assisting Retirees In Their Daily Lives

Abstract

The proportion of older adults is increasing in many countries, and many older adults have reduced work-related and social activities after retirement, which may pose negative challenges to their health. Recent advancements in AI, in particular Large Language Models (LLMs), might have the potential to enrich retirees’ daily experiences with their improved natural language understanding and conversation construction abilities. In this work, we sought to explore the potential ways in which retirees might interact with LLMs and their expectations by conducting participatory workshops and a three-week follow-up longitudinal study with 25 retirees to gain insight into their interaction behavior with and perception of LLMs, exemplified by ChatGPT. Our results reveal a range of ways in which older adults feel LLMs may be able to assist them in their daily lives and pinpoint the challenges and concerns faced by this group. Based on the findings, we discussed five key design implications to optimize LLMs-empowered interactions for assisting older adults.

Ähnliche Arbeiten