Social robots eased loneliness in older adults, a review of 19 studies found

The size of the effect depended on setting and study design. Studies with control groups showed larger reductions than studies that only compared before and after. People in nursing homes, dementia care centers and hospitals gained more than those living in their own homes, and effects were stronger in Japan and Turkey than in the United States.

Age, cognitive impairment, robot type (pet, humanoid or voice assistant), intervention length and publication year did not significantly change the results. The authors read this as evidence that the relationship with the robot matters more than its technical sophistication, and they call for equity, privacy and consent safeguards.

The authors used robust variance estimation to handle studies reporting several effect sizes, and calculated effects separately for controlled studies and single-group pre-post studies. Missing standard deviations, under 24% of the data, were filled in by predictive mean matching. Two reviewers screened and extracted data, with inter-rater agreement of κ = 0.79.

Robots in the included studies ranged from pet-like devices such as robotic cats to humanoids and voice assistants such as Amazon Alexa. Because robot type did not change results, the authors argue that even simple interventions may help, and that the meaning older adults give the interaction matters more than the robot's features.

On policy, the authors say social robots should complement rather than replace human contact. They warn that without equity-focused strategies the benefits may favor people with higher income or digital literacy, and that privacy, informed consent and fairness need attention for older adults with cognitive impairments.