Using an AI-Driven Child Chatbot Avatar as a Training Tool for Information Gathering Skills of Dental and Medical Professionals: A Pilot Study

Abstract

Healthcare professionals often receive limited training in information gathering, especially for cases of suspected child maltreatment. This pilot study evaluated a brief interview training program using an artificial intelligence-driven child avatar chatbot to simulate realistic encounters with children. GPT-3 and one-shot prompting were used to simulate an 8-year-old girl who was allegedly subjected to violence. A pretest-posttest design assessed the program's impact, with all participants (n = 18) receiving a video lecture and indirect feedback on their questions based on the avatar's responses. One group (n = 10) received direct feedback on their use of open and closed questions, along with examples after each interview. Results showed no significant improvement in interviewing behaviour, regardless of feedback condition. This study highlights the need for further development and tailoring of training programs to better support healthcare professionals in cases of suspected child abuse.