Psychology · Social Psychology
An AI model predicted suicidal thoughts and behaviors more accurately than traditional statistical methods among a group of 3,901 Taiwanese university students. The researchers say its interpretable design could support screening, but the study does not show whether using the model improves care or prevents suicide-related outcomes.
In a randomized study of 388 university students in China, an AI counseling agent improved perceived rapport and willingness to disclose. It recognized general risks better, but was less reliable with crisis-level disclosures, pointing to a supporting role alongside trained counselors.
A study of female university students living in dormitories in Baghdad links social media use to worse mental health, including more anxiety, depression, and social isolation. The researchers report that longer daily use and entertainment or social comparison content were associated with these mental health indicators.
A study in Scientific Reports tested an AI system that predicts suicide-related behaviour within 30 days of an emergency department visit using structured triage information available at admission. The best-performing model, a LightGBM system, showed AUROC 0.88 and good recall on an independent test set, and highlight likely drivers such as prior psychiatric history, age, and physiological variables.
A national U.S. survey found that 24% of adults use general-purpose AI chatbots for mental health. Researchers estimate that 14–18 million U.S. adults may use these tools for mental health in 2024, and that users report poorer mental health and greater difficulty accessing traditional care.
Generative AI chatbots are increasingly used by young people for mental health support, but their safety for suicide-risk situations is not well understood. In a review, researchers found that existing studies often test chatbots with simple, one-turn questions and that results worsen when risk is indirect or gradually revealed.
