Conversational AI chatbots can effectively reduce feelings of loneliness and social anxiety. (Image courtesy of UNIST)
ULSAN, Jan. 21 (Korea Bizwire) — A new study has revealed that conversational AI chatbots can effectively reduce feelings of loneliness and social anxiety, according to research findings released on January 20 by scientists at the Ulsan National Institute of Science and Technology (UNIST).
The research, conducted jointly with Korea University Anam Hospital’s Department of Psychiatry, examined the psychological impact of interactions with an AI chatbot named Iruda 2.0 on mental health outcomes.
In the four-week study, 176 participants engaged in conversations with the chatbot at least three times per week. Researchers measured participants’ loneliness and social anxiety levels using standardized questionnaires and conducted follow-up interviews to analyze their experiences in depth. The study focused on evaluating how chatbot interactions contributed to emotional stability and social connectedness.
The results were striking: regular interactions with the chatbot led to an average 15% reduction in loneliness scores and an 18% decrease in social anxiety measurements. The benefits were particularly pronounced among users who shared more personal information about their emotions, thoughts, and experiences with the chatbot, as well as those with higher resilience levels.
“The study demonstrates that chatbots can serve as effective digital tools for alleviating loneliness and anxiety,” said Kim Myung-sung, the study’s lead author and a doctoral candidate at UNIST’s Graduate School of Health Science. “We’ve empirically shown that these systems can provide emotional support beyond mere technological functionality, and that interpersonal factors significantly influence their effectiveness.”
Jung Dooyoung, a professor of biomedical engineering at UNIST, emphasized the potential applications: “When used safely, these systems could be particularly valuable in preventing mental health issues in environments where professional mental health resources are limited.”
The research team plans to conduct additional studies focused on improving the chatbot’s usability and personalized services. Their findings were published online in the Journal of Medical Internet Research (JMIR) on January 14.
Kevin Lee (kevinlee@koreabizwire.com)
