"Though Helpful, Still Hesitant": A TAM-Based Qualitative Study on Older Adults' Ambivalent Acceptance and Model Extensions in AI Fitness Coaches

YI YAU^*Ya Chun Shen

• Taipei City University of Science and Technology, Taipei, Taiwan

Background: With the deepening integration of artificial intelligence (AI) and health promotion strategies, digital health applications have continued to expand. AI fitness coaches have emerged as promising tools to support physical activity among older adults. Although models such as the Unified Theory of Acceptance and Use of Technology and the Senior Technology Acceptance Model have extended the Technology Acceptance Model (TAM) by introducing external factors across technological, psychological, and socio-emotional domains, further research is needed to contextualize these variables within AI-based environments. Prior studies have focused mainly on general digital or assistive technologies, with limited attention to emotionally and cognitively complex human–AI interactions such as AI fitness coaching. Therefore, this study adopts TAM as the core theoretical framework, as its concise structure suits exploratory interviews while preserving participants’ lived experience richness.Methods: Participants were recruited through public announcements at sports and neighborhood activity centers in Northern Taiwan, with support from community leaders. This qualitative study used TAM as a foundational framework and conducted semi-structured interviews with 12 older adult participants. The interview protocol explored their interactions with AI fitness coaches, emphasizing perceived usefulness, ease of use, emotional responses, and contextual influences. Verbatim transcripts were thematically analyzed using both deductive codes from TAM and inductive codes to capture emerging themes. Two researchers independently coded the data and reached consensus to ensure analytic rigor.Results: Perceived usefulness and ease of use were shaped not only by system design but also by aging-related factors such as physical limitations, life history, and autonomy. While participants expressed cognitive approval of AI fitness coaches, emotional resistance revealed ambivalent acceptance. Findings highlight that emotional, social, and cultural factors—often overlooked in TAM—play a significant role in shaping older adults’ technology engagement.Conclusion: This study expands TAM by emphasizing that technology acceptance is a dynamic, context-sensitive process shaped by emotional and cultural influences. It underscores the need to adapt external variables to specific settings and encourages designing AI fitness technologies that foster emotional resonance and cultural relevance. The concept of "ambivalent acceptance" offers new insight into how aging individuals engage with technology, contributing to both theoretical development and inclusive design.