AI Companions Could Diminish Social Skills in Isolated Users, Study Finds

Artificial intelligence companions are rapidly moving from niche experiments to everyday use, prompting a new study by Singapore Management University and Duke‑NUS Medical School that shifts the conversation from “beneficial or harmful” to “who gains and who is left behind.” The researchers argue that as these digital partners become embedded in homes, workplaces and healthcare settings, the real policy challenge is ensuring that the advantages of AI—such as personalized support, mental‑health monitoring and elder care—are distributed fairly across socioeconomic, ethnic and age groups.

The study maps out the potential risks for vulnerable populations, including reduced access for low‑income users, privacy concerns for marginalized communities, and the possibility of reinforcing existing health disparities. It recommends a multi‑layered approach: transparent data governance, inclusive design processes that involve diverse user groups, and public funding mechanisms to subsidize high‑quality AI services for underserved sectors. By foregrounding equity, the authors suggest that society can steer the rise of AI companions toward a future where technological benefits are shared broadly rather than concentrated in a privileged few.