Baraccani, Daniela
(2025)
Artificial Intelligence in child abuse and neglect: a pilot study on ethical constraints and opportunities.
[Laurea magistrale], Università di Bologna, Corso di Studio in
Artificial intelligence [LM-DM270]
Documenti full-text disponibili:
Abstract
According to the latest national investigation carried out by the Guarantor Authority for Childhood and Adolescence (Autorità Garante per l’Infanzia e l’Adolescenza) the cases of maltreatment of children and adolescents have increased by 58% in five years. The need to have an effective tool and system for the early identification of abuse and maltreatment led Sant’Orsola’s doctors to propose a collaboration with the Artificial Intelligence department of the University of Bologna to study the possibility of including AI technology to tackle this problem.
In this thesis, after researching known solutions, we performed rule-extraction on the professional manuals provided by the doctors. The manuals describe the symptoms that can be found in the patients at risk, along with the behavior that the doctor should have during the patient-doctor interaction and the steps that the doctor and the hospital should follow.
After extracting the rules, we analyzed, restructured and summarized them into questions. We studied two important instruments, ESCAPE and SCAN, that were born with the same aim as our project and compared our questions with the questionnaires described in these two studies.
This project resulted in the creation of a 5-item checklist that is designed to assist the medical professional that encounters an at-risk patient to determine whether the child could be a victim of abuse or maltreatment.
In the future, the aim is to incorporate this tool as a standard step during any clinical consultation of a minor, in order to drastically reduce the cases of maltreatment and abuse of children and adolescents in Italy.
Abstract
According to the latest national investigation carried out by the Guarantor Authority for Childhood and Adolescence (Autorità Garante per l’Infanzia e l’Adolescenza) the cases of maltreatment of children and adolescents have increased by 58% in five years. The need to have an effective tool and system for the early identification of abuse and maltreatment led Sant’Orsola’s doctors to propose a collaboration with the Artificial Intelligence department of the University of Bologna to study the possibility of including AI technology to tackle this problem.
In this thesis, after researching known solutions, we performed rule-extraction on the professional manuals provided by the doctors. The manuals describe the symptoms that can be found in the patients at risk, along with the behavior that the doctor should have during the patient-doctor interaction and the steps that the doctor and the hospital should follow.
After extracting the rules, we analyzed, restructured and summarized them into questions. We studied two important instruments, ESCAPE and SCAN, that were born with the same aim as our project and compared our questions with the questionnaires described in these two studies.
This project resulted in the creation of a 5-item checklist that is designed to assist the medical professional that encounters an at-risk patient to determine whether the child could be a victim of abuse or maltreatment.
In the future, the aim is to incorporate this tool as a standard step during any clinical consultation of a minor, in order to drastically reduce the cases of maltreatment and abuse of children and adolescents in Italy.
Tipologia del documento
Tesi di laurea
(Laurea magistrale)
Autore della tesi
Baraccani, Daniela
Relatore della tesi
Correlatore della tesi
Scuola
Corso di studio
Ordinamento Cds
DM270
Parole chiave
rule-extraction, llm, ai, child abuse, maltreatment, pediatrics, artificial intelligence
Data di discussione della Tesi
22 Luglio 2025
URI
Altri metadati
Tipologia del documento
Tesi di laurea
(NON SPECIFICATO)
Autore della tesi
Baraccani, Daniela
Relatore della tesi
Correlatore della tesi
Scuola
Corso di studio
Ordinamento Cds
DM270
Parole chiave
rule-extraction, llm, ai, child abuse, maltreatment, pediatrics, artificial intelligence
Data di discussione della Tesi
22 Luglio 2025
URI
Statistica sui download
Gestione del documento: