A groundbreaking development in mental health assessment has emerged as a powerful ally in the fight against self-harm and suicide. Researchers from Northwestern University, the University of Cincinnati, Aristotle University of Thessaloniki, and Massachusetts General Hospital at Harvard School of Medicine have unveiled a new assessment tool powered by artificial intelligence (AI) that can predict suicidal thoughts and behaviours with an impressive 92% accuracy.

The innovative system, detailed in a study published in Nature Mental Health, relies on a simple picture-ranking task combined with a small set of contextual and demographic variables. Participants, aged 18 to 70, were asked to rank a sequence of 48 pictures across various categories on a scale of like-to-dislike. Additionally, demographic information such as age, gender, education level, and handedness was collected.

Lead author Shamal Shashi Lalvani, a doctoral student at Northwestern University, highlighted the significance of the tool's ability to quantify human behaviour and preferences in predicting suicidality. He emphasised the potential of this approach to provide a more quantitative understanding of mental health, drawing connections to disciplines like behavioural economics.

Hans Breiter, a professor at the University of Cincinnati and contact person for the study, underscored the urgent need for such technology, particularly in addressing the alarming rates of self-harm among veterans and students in the U.S. Breiter expressed optimism about the tool's potential to improve outcomes and alleviate the strain on the healthcare system.

Co-senior author Aggelos Katsaggelos, from Northwestern University, highlighted the simplicity yet effectiveness of the picture-rating task in predicting suicidality. He noted that the results exceeded typical methods of measurement without relying on extensive electronic health records or big data.

The AI system developed by Northwestern and the University of Cincinnati proved capable of predicting four measures related to suicidal thoughts and behaviours, including passive and active ideation, planning for suicide, and coping strategies. However, researchers cautioned about potential cultural influences on prediction success and the self-reported nature of the surveys.

Despite these limitations, the new AI tool represents a significant step forward in mental health assessment, offering hope for more accurate identification and intervention for individuals at risk of self-harm. As rates of loneliness and self-harm continue to rise, innovative solutions like this AI-driven assessment tool hold promise for saving lives and improving mental well-being.