Abstract
Background: With global aging, robots are considered a promising solution for handling the shortage of aged care and companionships. However, these technologies would serve little purpose if their intended users do not accept them. While the socioemotional selectivity theory predicts that older adults would accept robots that offer emotionally meaningful relationships, selective optimization with compensation model predicts that older adults would accept robots that compensate for their functional losses. Objective: The present study aims to understand older adults’ expectations for robots and to compare older adults’ acceptance ratings for 2 existing robots: one of them is a more human-like and more service-oriented robot and the other one is a more animal-like and more companion-oriented robot. Methods: A mixed methods study was conducted with 33 healthy, community-dwelling Taiwanese older adults (age range: 59–82 years). Participants first completed a semi-structured interview regarding their ideal robot. After receiving information about the 2 existing robots, they then completed the Unified Theory of Acceptance and Use of Technology questionnaires to report their pre-implementation acceptance of the 2 robots. Results: Interviews were transcribed for conventional content analysis with satisfactory inter-rater reliability. From the interview data, a collection of older adults’ ideal robot characteristics emerged with highlights of humanlike qualities. From the questionnaire data, respondents showed a higher level of acceptance toward the more service-oriented robot than the more companion-oriented robot in terms of attitude, perceived adaptiveness, and perceived usefulness. From the mixed methods analyses, the finding that older adults had a higher level of positive attitude towards the more service-oriented robot than the more companion-oriented robot was predicted by higher expectation or preference for robots with more service-related functions. Conclusion: This study identified older adults’ preference toward more functional and humanlike robots. Our findings provide practical suggestions for future robot designs that target the older population.
