This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: As AI companions become more commonplace, new research from Singapore Management University (SMU) and Duke-NUS Medical School argues that the real question is no longer whether these technologies are beneficial or harmful, but who benefits, who is most at risk and how society can ensure AI serves everyone equitably. Published recently in Nature Human Behaviour, the paper, titled "How AI companions could deepen social inequality," introduced a new framework for understanding how AI companions may shape existing social inequalities.
The study was led by first author Zhang Qiyang, an assistant professor in learning analytics at SMU College of Integrative Studies; Zhang Renwen, a Nanyang assistant professor from the Wee Kim Wee School of Communication and Information at Nanyang Technological University; and senior author Liu Nan, an associate professor at the Centre for Biomedical Data Science at Duke-NUS Medical School. Rather than viewing AI companions through a simple "good or bad" lens, the researchers examined how differences in users' social support, technology design and governance can lead to uneven outcomes. They also outlined practical recommendations to help policymakers, technology developers and users maximize the technology's benefits while mitigating its risks.
"The debate around AI companions has largely focused on whether the technology is beneficial or harmful. Our research shows that this is the wrong question," Zhang Qiyang said. "The more important question is who benefits, who bears the risks, and why those outcomes are so unevenly distributed.
Without deliberate safeguards, AI companions risk reinforcing existing social inequalities, leaving those who most need human support the most exposed to harm." The study introduced a new framework for understanding how AI companions may reinforce existing inequalities across three interconnected dimensions: users, platform design and governance. Rather than viewing risks in isolation, the framework showed how these factors interact to shape unequal outcomes and proposed practical interventions at each level. The study unveiled a key insight: a "rich-get-richer" dynamic in human relationships.
People with strong family ties and social networks are more likely to use AI companions to supplement existing relationships, such as rehearsing difficult conversations or managing stress. In contrast, those who are lonely, socially isolated or have limited access to mental health support are more likely to rely on AI companions as a substitute for human connection. Over time, this reliance may contribute to social deskilling—the gradual erosion of interpersonal skills through reduced opportunities to practice authentic human interaction.
To explain how these inequalities emerge, the study adapted the Swiss cheese model, a framework widely used in engineering and safety science to explain how multiple small failures align to produce larger systemic risks. Instead of viewing AI companionship as a problem caused by any single factor, the study argued that harm occurs when several protective layers fail simultaneously. These included users' AI literacy, social support networks, platform design choices and regulatory oversight.
Extract — continue reading at the source.
