Artificial Intelligence Succeeds in Predicting Children's Mental Health Issues by Analyzing Linguistic Styles

A recent psychological study conducted by researchers at a US university revealed the possibility of using artificial intelligence models to predict future mental health issues in children by analyzing their natural everyday language. The study involved over 200 children aged between 9 and 13, where the models demonstrated high accuracy in determining the likelihood of them developing mental disorders later on. The researchers concluded that sentence structure style and the use of short conjunctions were stronger indicators of declining mental health than the content of the speech itself. The study emphasized that early prediction prior to adolescence is crucial, given the difficulty of treating depression and anxiety once they are established, and the lack of clear prior symptoms. The children were assessed by analyzing audio recordings of interviews covering various topics and unfortunate events, alongside expert questionnaires measuring the extent of stress and traumatic experiences, which ranged from economic hardship and divorce to abuse. The results showed that using expressions referring to physical violence or social exclusion was one of the strongest predictors of mental health issues, whereas phrases related to psychological support, friendship, and social relationships were linked to sound mental health.