AI Avatar Tells You What Happened: The First Test of Using AI-Operated Children in Simulated Interviews to Train Investigative Interviewers

Abstract

Background: Previous research has shown that simulated child sexual abuse interview training using avatars paired with feedback and modeling improves interview quality. However, in order to make this approach scalable, the classification of interviewer questions needs to be automated.Objective: We tested an automated question classification system for simulated investigative interviews with children while also providing interventions (feedback and modeling) aimed at improving interview quality. Participants and Setting: Forty-two professionals were randomly assigned to the no intervention, feedback, or modeling group. Methods: The participants conducted two simulated child sexual abuse interviews online while receiving no intervention, feedback or modeling after the first interview. Feedback consisted of the outcome of the alleged case and comments on the quality of the questions asked by the interviewer. Modeling consisted of learning points and videos illustrating good and bad questioning methods.Results: The total percentage of agreement in question coding between human operators and the automated classification was 72% for the main categories (recommended vs. not recommended) and 52% when the eleven subcategories were considered. The intervention groups improved from first to second interview while this was not the case in the no intervention group (intervention x time: p = .007, ηp2 = .28).Conclusions: Automated question classification worked well for classifying the interviewers’ questions allowing interventions to improve the quality of the interviews.

Ähnliche Arbeiten