Large Language Model (LLM) and Human Performance in Child Investigative Interviewing Question Formulation Tasks
Behavioral Sciences & the Law · 2026-02 · 44 (1) · 142-163
DOI: 10.1002/bsl.70029 ↗Kurz gefasst
Fachartikel in der Zeitschrift Behavioral Sciences & the Law, online erschienen am 08.12.2025. Verfasst von Liisa Järvilehto und 6 weiteren Autor:innen. Beteiligte Einrichtung: Åbo Akademi University, Turku, Finland. Hauptthemen laut MEDLINE: Language, Interviews as Topic. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 57 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Sundell Foundation.
Abstract (Originalsprache)
We compared the performance of large language models (LLMs) and humans with various levels of expertise in child investigative interviewing on tasks related to question formulation. Two tasks were employed: a static Interview Excerpt Task where participants (60 psychologists, 60 naive participants, GPT‐4, and Llama‐2) formulated follow‐up questions to 100 interview excerpts, and a dynamic Avatar Interviewing Task where participants (32 professionals, 32 students, and GPT‐4) conducted 10‐min interviews with AI‐driven child avatars. In the dynamic task, LLMs used fewer recommended questions ( M = 8.69 vs. 18.75) and more non‐recommended questions ( M = 17.69 vs. 6.81) than professionals. Conversely, in the static task, GPT‐4 outperformed psychologists, using more invitations (67.8% vs. 5.4%) and fewer option‐posing questions (3.7% vs. 31.4%). While LLMs demonstrated strong question formulation skills in controlled environments, they struggled with adaptive dialogs.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Åbo Akademi University, Turku, Finland.
- Forensic Psychology Center for Children and Adolescents, Helsinki University Hospital, Helsinki, Finland.
- East China Normal University, Shanghai, China.
- New York University Shanghai, Shanghai, China.
- Faculty of Psychology, Meiji Gakuin University, Tokyo, Japan.
- European Institute for Crime Prevention and Control, Affiliated with the United Nations (HEUNI), Helsinki, Finland.
Förderung
- Sundell Foundation
Zitieren
APA
Järvilehto, L., Sun, Y., Aiba, N., Haginoya, S., Hallström, H., Korkman, J., & Santtila, P. (2026). Large Language Model (LLM) and Human Performance in Child Investigative Interviewing Question Formulation Tasks. Behavioral Sciences & the Law, 44(1), 142-163. https://doi.org/10.1002/bsl.70029
Vancouver
Järvilehto L, Sun Y, Aiba N, Haginoya S, Hallström H, Korkman J, et al. Large Language Model (LLM) and Human Performance in Child Investigative Interviewing Question Formulation Tasks. Behavioral Sciences & the Law. 2026;44(1):142-163. doi:10.1002/bsl.70029
BibTeX
@article{Jrvilehto2026,
author = {Järvilehto, Liisa and Sun, Yongjie and Aiba, Nami and Haginoya, Shumpei and Hallström, Hasse and Korkman, Julia and Santtila, Pekka},
title = {Large Language Model (LLM) and Human Performance in Child Investigative Interviewing Question Formulation Tasks},
journal = {Behavioral Sciences & the Law},
year = {2026},
volume = {44},
number = {1},
pages = {142--163},
doi = {10.1002/bsl.70029}
}
EUVIMED