A new study from Singapore Management University (SMU) and Duke-NUS Medical School suggests that AI companions—virtual entities designed to interact with humans—could deepen social inequalities. The research, published in Nature Human Behaviour, explores how these AI systems affect different users, platform design, and governance structures. It was led by Zhang Qiyang, an assistant professor at SMU, along with researchers from Nanyang Technological University and Duke-NUS Medical School. The study proposes a framework to examine how differences in users' access to social support, the design of the technology, and regulatory oversight can lead to uneven outcomes across society. The research argues that discussions about AI companions should move beyond simple questions of whether they are good or bad. Instead, the focus should be on understanding who benefits from these technologies, who might be harmed, and how society can ensure fair use. The study highlights a "rich-get-richer" dynamic in human relationships, where individuals with strong social networks use AI companions to enhance existing relationships, while those who are socially isolated may turn to AI as a substitute for human interaction. This could lead to a gradual decline in interpersonal skills over time. To explain how these risks might emerge, the researchers adapted the "Swiss cheese model," a concept from engineering and safety science that shows how multiple small failures can align to create larger problems. They found that governance is the most pressing concern, as AI companions currently exist in a regulatory gray area in many countries. The researchers suggest that treating AI companions as health-related technologies could lead to stronger protections, such as ensuring they are designed with age-appropriate features, providing clear information to users, and including protocols for responding to crises. Singapore is highlighted in the study as a country that is both well-positioned and uniquely vulnerable to the rise of AI companions. With its advanced digital infrastructure and an aging population, Singapore faces both opportunities and challenges in managing the impact of these technologies. The study recommends practical steps to ensure AI companions contribute positively to social well-being, including recognizing their potential effects on mental health and adopting a layered governance approach that combines regulation, public education on AI, and responsible platform design.