A team of US researchers announced the development of a new AI-based tool that can predict suicidal thoughts and behaviors instantly. The tool, which involves a simple picture-ranking task and a small set of contextual/demographic variables, was found to be 92% effective on average. Researchers from Northwestern University, University of Cincinnati, Aristotle University, and Massachusetts General Hospital/Harvard School of Medicine collaborated on this project.
The first author, doctoral student Shamal Shashi Lalvani from Northwestern University, highlighted the significance of a system that assesses reward and aversion judgments to understand preference behavior and predict suicidality using interpretable variables describing human behavior.
The study, published in the journal Nature Mental Health, suggested that the tool could aid medical professionals, hospitals, and the military in identifying individuals at high risk of self-harm. The research, based on a survey of 4,019 individuals aged 18 to 70 in the US, demonstrated the software's ability to predict various aspects of suicidal behavior.