Dialogue Agents that Share Family Information to Strengthen Grandparent-Grandchild Relationships

A team from Osaka University, led by Seiya Mitsuno, has published research on a novel AI dialogue agent designed to combat social isolation in older adults by strengthening intergenerational bonds. Unlike standard chatbots that simulate companionship, this agent actively facilitates real-world relationships. It operates by engaging in daily conversations with both grandparents and their grandchildren on a chatbot platform, then strategically sharing non-sensitive, everyday information gathered from each party to spark more meaningful engagement between the actual family members.

The researchers conducted a rigorous 10-day empirical experiment involving 108 grandparent-grandchild pairs to test the system's efficacy. The findings were significant: older adults showed increased willingness to interact with the agent specifically because it shared information about their grandchildren. More importantly, the study measured a strengthened psychological connection between the grandparents and grandchildren themselves. Furthermore, daily interactions with the agent were correlated with reduced anxiety levels in both groups, highlighting a dual mental health benefit.

This work moves beyond treating AI as a standalone companion and instead positions it as a relational catalyst. The study provides concrete design insights for creating agents that address core social needs, suggesting the most effective support for isolation may come from technology that nurtures users' existing human networks rather than attempting to replace them.