Imagine a personal AI agent that you can converse with naturally about your friends and family, helping you remember previous topics you’ve discussed, suggesting topics for future conversations, reminding you of prior commitments made with others, and encouraging you to stay connected with your relations. To engineer such a system, one would need to consider how to process conversational data to distill important information, how to identify which threads of conversation are related to which relation of interest, and how to integrate this information into natural conversation with a human user. I hope to develop this system, in hopes of both advancing the area of organic, fluid human-AI interaction, and in hopes of addressing the rising issue of loneliness.
This system could be especially useful for older adults, who are becoming increasingly socially isolated due to various factors including disability-related barriers and their lower digital literacy in using social media and other forms of communication technology [1]. Loneliness, the subjective experience of isolation, has subsequently become more prevalent; this state has been shown to lead to poor health outcomes comparable to well-known health risk factors such as smoking and alcohol consumption [2].
In my research I explore whether conversational AI characters could offer a solution for this social epidemic—specifically, by strengthening an individual’s social circle through AI agents.