Loneliness is a prevalent issue affecting individuals of all age groups, and finding effective solutions to combat it is crucial for public health. Recent advancements in technology have introduced the concept of ‘relational agents’, which are software agents powered by artificial intelligence and large language models (LLMs). A systematic review and meta-analysis were conducted to assess the efficacy of relational agents in alleviating loneliness across different age demographics.

Published on July 6, 2024, in BMC Public Health, the study delved into various interventions aimed at addressing loneliness, such as enhancing social skills, social support, social interaction, and challenging maladaptive thoughts. The research team searched through 11 databases, including Ovid MEDLINE and Embase, from their inception to September 16, 2022, to gather relevant studies for analysis.

Out of 3,935 records initially identified, 14 studies met the inclusion criteria and were subjected to meta-analysis. These studies encompassed a total of 286 participants, with individual sample sizes ranging from 4 to 42 individuals. Statistical analyses revealed a significant reduction in loneliness levels among participants who interacted with relational agents.

The meta-analysis employed a random-effects model to calculate pooled estimates of Hedge’s g, with additional sensitivity and subgroup analyses conducted to explore the data further. Evaluation for publication bias was carried out using funnel plots, Egger’s test, and a trim-and-fill algorithm to ensure the robustness of the findings.

This comprehensive investigation sheds light on the promising role of relational agents in addressing loneliness across diverse age groups. By leveraging the capabilities of artificial intelligence and LLMs, these software agents offer a novel and potentially scalable solution to tackle the pervasive issue of loneliness in society.