JMIR Publications

ABSTRACT

Background:

The older adult population is increasing rapidly and suffers from a multitude of chronic conditions affecting quality of life. Physical activity and reduced sedentary behaviour reduce adverse health conditions, and help preserve functional independence and cognitive abilities but the older adult age group has much lower than recommended levels of physical activity than other population demographics.

Objective:

The study aims to understand older adults’ motivations towards health and physical activity, the role of technology in achieving their health and physical activity goals, and explore the attitudes towards use of generative artificial intelligence (GenAI) in health applications.

Methods:

This study is qualitative and based on hourlong interviews with 20 older adults based in the United States. To elicit specific design preferences, participants were presented with visual probes of AI-generated avatars. The interviews were analysed using thematic analysis, and the coding process drew on grounded theory techniques.

Results:

Older adults primarily wished to maintain functional independence: both physical and cognitive. They reported satisfaction with levels of physical activity adjusting them to chronic conditions and injuries. They were active through everyday functions of living as well as structured exercise programs with many using wearable devices and smartphones to help monitor their activity. They reported being motivated by social accountability as well as gamification. Accessibility issues were mentioned frequently including problems with touchscreens, limited viewing, difficulty navigating interfaces with clutter, complex multitouch processes for undertaking actions, and steep learning curves; computers were preferred for certain tasks over mobile phones due to easier viewability, keyboard access, and greater security features. Personalisation was highly desired in health apps. There was a willingness to use more health technology, including artificial intelligence (AI) and they were willing to consider AI in both motivational and informational roles in applications as long as it was personalised. AI embodiment preferences varied widely but with an emphasis on relatability. However, there was marked scepticism towards data privacy, concerns about loss of independence and self-efficacy, and doubts about AI’s accuracy.

Conclusions:

Accessibility barriers still exist in apps. Social accountability is important for motivation. AI has some potential in meeting accountability and informational needs especially with the use of gamification. Future use of GenAI in health for older adults needs to focus on personalisation, accuracy and detailed referencing of health information. Some recommendations for health application design for physical activity are provided.