Bullying and cyberbullying are widespread issues that profoundly affect the mental health and overall well-being of young people. To address these matters, we propose an AI-powered tool to group and tag individuals based on their responses to profiling surveys, providing educators with insights into (cyber)bullying dynamics across a community of students. This research is conducted within the scope of the BullyBuster (BB) project [5], which aims to combine AI, law and psychology to develop a comprehensive framework for the prevention and detection of bullying or cyberbullying. The framework includes several modules for a range of complementary objectives, such as spotting physical violence, deepfake recognition and behavioural modelling of subjects based on surveys. The latter, in particular, were the foundation for the current analysis. Given the sensitive nature of the matter, the tool is built upon the principles of human-centered AI, prioritizing transparency and explainability of each output, hence informed decisions for an effective and timely intervention.
Explainable AI for Bullying and Cyberbullying Detection / Santoro, Enrico; Marrone, Stefano; Sansone, Carlo. - (2024), pp. 59-59. (Intervento presentato al convegno 2024 Conference on Human Centred Artificial Intelligence - Education and Practice) [10.1145/3701268.3701289].
Explainable AI for Bullying and Cyberbullying Detection
Santoro, Enrico;Marrone, Stefano;Sansone, Carlo
2024
Abstract
Bullying and cyberbullying are widespread issues that profoundly affect the mental health and overall well-being of young people. To address these matters, we propose an AI-powered tool to group and tag individuals based on their responses to profiling surveys, providing educators with insights into (cyber)bullying dynamics across a community of students. This research is conducted within the scope of the BullyBuster (BB) project [5], which aims to combine AI, law and psychology to develop a comprehensive framework for the prevention and detection of bullying or cyberbullying. The framework includes several modules for a range of complementary objectives, such as spotting physical violence, deepfake recognition and behavioural modelling of subjects based on surveys. The latter, in particular, were the foundation for the current analysis. Given the sensitive nature of the matter, the tool is built upon the principles of human-centered AI, prioritizing transparency and explainability of each output, hence informed decisions for an effective and timely intervention.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.