Co-Designing a Human–AI Collaborative Support System for Independent Living Among Adults with Intellectual and Developmental Disabilities In South Korea: A Living Lab Pilot Study

In line with the United Nations Convention on the Rights of Persons with Disabilities, there has been a growing emphasis on community-based independent living for individuals with intellectual and developmental disabilities (IDD). In South Korea, independent living is largely supported through a national personal assistance system, resulting in a high reliance on Direct Support Professionals (DSPs). Due to the characteristics of IDD, individuals often require continuous and repetitive support in daily life. However, such demands are difficult to meet solely through human resources, given workforce shortages, high turnover, and significant emotional burden among DSPs. These structural constraints limit the extent to which individualized needs are fully addressed and may create gaps in daily interaction and emotional support.

Artificial intelligence (AI) has emerged as a promising complement to human support by assisting with repetitive tasks and providing consistent, individualized interactions. In particular, AI chatbots may support daily routines while also offering conversational engagement that can help alleviate loneliness. Human–AI collaboration may therefore not only enhance user outcomes but also improve service quality and working conditions for DSPs. This study aims to develop and examine the feasibility of a human–AI collaborative support system using a Living Lab approach in South Korea.

This pilot study employed a user-centered Living Lab methodology involving multi-stakeholder co-creation over 14 sessions. Participants included adults with IDD living independently, a service provider, an AI developer, and researchers. The study proceeded through four phases: needs identification, problem definition, prototype development, and evaluation. Data were collected through interviews, focus groups, and observations, and iteratively reviewed with stakeholders to identify key themes. An AI chatbot prototype (“Hi Buddy”) was developed through three cycles of co-design and refinement. Usability and feasibility were assessed based on ease of use, satisfaction, emotional response, and intention for continued use.

Participants identified unmet needs in communication, social interaction, and structured daily routines, as well as challenges related to loneliness and limited support availability. “Hi Buddy” supported daily activities such as meal preparation, scheduling, exercise, and leisure engagement. Both individuals with IDD and DSPs reported high usability and satisfaction. Users experienced improved daily structure and perceived autonomy, and some reported reduced loneliness through chatbot interaction. DSPs reported reduced repetitive workload and emotional burden, as well as improvements in service efficiency and consistency. These findings suggest that human–AI collaboration can enhance both user experiences and service delivery processes.

Findings suggest that a co-designed human–AI collaborative system is feasible and acceptable for supporting independent living in the Korean context. By complementing human support and providing consistent interaction, AI may enhance person-centered care while improving service quality and working conditions for DSPs. Although context-specific, this study offers broader implications for integrating human–AI collaboration into disability services. Future research should examine effectiveness using larger samples and longitudinal designs.