노인 맞춤형 ChatGPT 대화가 노인 대화 경험에 미치는 영향: 탐색적 연구 = An Exploratory Study on the Effects of a ChatGPT-Based Conversational System Designed for Older Adults on Their Communication Experience

다국어 초록 (Multilingual Abstract)

Purpose: This study aimed to explore the potential of supporting communication among older adults using an artificial intelligence-based conversational system, as well as to alleviate digital exclusion in the context of an increasingly aging society. In particular, this study focused on the possibility that interaction participation may be limited when generative artificial intelligence does not sufficiently reflect the cognitive and communicative characteristics of older adults, and examined the effects of applying an older adult- tailored communication guideline.
Methods: To this end, conversations with ChatGPT were conducted under guideline- applied and non- applied conditions with ten female older adults aged 65 years and above, and comparisons were made in terms of conversation satisfaction, verbal productivity, and interaction structure.
Results: The results showed no statistically significant differences in conversation satisfaction and verbal productivity; however, qualitative changes were observed in speech functions and discourse structure. Under the guideline- applied condition, older adults’ utterances extended beyond simple responses and developed into expanded discourse incorporating specific personal experiences and emotions. In addition, participants demonstrated a tendency to lead and sustain conversations with ChatGPT by utilizing contextualized emotional expressions.
Conclusion: The application of the elderly- tailored communication guidelines was found to reduce cognitive burden among older adults while simultaneously enhancing the structural expansion of discourse and the continuity of interaction. Furthermore, these findings suggest that such guidelines hold practical potential to promote interactional participation and facilitate discourse expansion among older adults within generative artificial intelligence environments. Therefore, this study confirmed that the proposed guidelines could serve as foundational data for design of age- friendly communication strategies aimed at mitigating the digital divide.