{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,1]],"date-time":"2026-08-01T00:44:06Z","timestamp":1785545046700,"version":"3.56.0"},"reference-count":81,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec sec-type=\"background\">\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Child abuse severely disrupts the healthy growth and development of children, resulting in long-term physical as well as emotional consequences. In real-world settings, despite the continuous increase in reported abuse cases, the chronic shortage of certified child abuse professionals has significantly increased the workload of individual counselors, making timely intervention increasingly difficult.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"objective\">\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>This study aims to reduce the workload of counselors in real-world counseling settings by supporting counseling and child abuse detection processes. We propose a conversational artificial intelligence\u2013based framework, Conversational Artificial Intelligence for Child Abuse Detection (CACAD), which conducts counseling with children and detects child abuse during the counseling process.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"methods\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>CACAD uses a large language model (LLM) to conduct counseling and to detect 4 types of child abuse: neglect, emotional, physical, and sexual. During the question generation process, the LLM serves as the primary agent, supported by 2 auxiliary modules. In the process of counseling, the LLM first determines whether a child\u2019s response provides sufficient information to understand the current situation. If the response is deemed insufficient, the LLM generates follow-up questions to elicit additional information while preserving the context of the previous question. Once sufficient information is obtained, the next question category prediction module predicts the most appropriate category for the subsequent question and passes it to the LLM, enabling more flexible guidance of the counseling flow. In parallel, the abusive question detection module filters out potentially harmful or inappropriate questions to protect children. For abuse detection, CACAD uses an instruction-tuned LLM specialized for child abuse detection, and uncertainty quantification is applied to dynamically flag cases as pending review for counselor confirmation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"results\">\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Experimental results using a Korean child-and-adolescent counseling dataset show that CACAD achieves strong performance in child abuse detection, with an exact match of 0.907 and a macro\n                      <jats:italic>\u2013F<\/jats:italic>\n                      <jats:sub>1<\/jats:sub>\n                      -score of 0.939. In addition, CACAD demonstrates effective performance in counseling-related tasks, including next question category prediction and abusive question detection, contributing to coherent and safe counseling interactions. Human evaluation by domain experts further confirms the reliability of CACAD in counseling sessions, and uncertainty-based selective prediction allows the system to dynamically identify cases requiring human review.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"conclusions\">\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>These findings demonstrate that LLM-based conversational agents can reliably perform child abuse detection within counseling conversations in real-world settings. The results further indicate that such systems can support counseling processes by integrating abuse detection, safe question generation, and uncertainty-aware decision handling in a unified framework.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.2196\/86536","type":"journal-article","created":{"date-parts":[[2026,5,20]],"date-time":"2026-05-20T14:25:06Z","timestamp":1779287106000},"page":"e86536-e86536","source":"Crossref","is-referenced-by-count":0,"title":["Conversational AI for Child Abuse Detection Through Multistage Counseling: Model Development and Validation Study"],"prefix":"10.2196","volume":"28","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-6249-4317","authenticated-orcid":false,"given":"Hyun-Young","family":"Moon","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-3798-7191","authenticated-orcid":false,"given":"Youn-Gyu","family":"Jin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-2479-544X","authenticated-orcid":false,"given":"YoonJu","family":"Kim","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-1733-7092","authenticated-orcid":false,"given":"Gwang-Cheol","family":"Lee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-8416-0564","authenticated-orcid":false,"given":"Hyeontaek","family":"Oh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-0035-6452","authenticated-orcid":false,"given":"Hyun A","family":"Kim","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-4133-1379","authenticated-orcid":false,"given":"Dinara","family":"Aliyeva","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6407-7427","authenticated-orcid":false,"given":"Hyunjoo","family":"Na","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2335-7072","authenticated-orcid":false,"given":"Kang-Min","family":"Kim","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,7,10]]},"reference":[{"key":"R1","unstructured":"Peace, dignity and equality on a healthy planet. United Nations. URL: https:\/\/www.un.org\/en\/ [Accessed 12-10-2025]"},{"key":"R2","unstructured":"Ebola outbreak in Africa. World Health Organization. URL: https:\/\/www.who.int\/ [Accessed 12-10-2025]"},{"key":"R3","unstructured":"UNICEF for every child. UNICEF. URL: https:\/\/www.unicef.org\/ [Accessed 12-10-2025]"},{"key":"R4","unstructured":"Global status report on preventing violence against children 2020. World Health Organization; 2020. URL: https:\/\/www.who.int\/teams\/social-determinants-of-health\/violence-prevention\/global-status-report-on-violence-against-children-2020 [Accessed 18-06-2026]"},{"issue":"4","key":"R5","doi-asserted-by":"publisher","first-page":"245","DOI":"10.1016\/s0749-3797(98)00017-8","article-title":"Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study","volume":"14","author":"Felitti","journal-title":"Am J Prev Med"},{"issue":"10","key":"R6","doi-asserted-by":"publisher","first-page":"652","DOI":"10.1038\/nrn.2016.111","article-title":"The effects of childhood maltreatment on brain structure, function and connectivity","volume":"17","author":"Teicher","journal-title":"Nat Rev Neurosci"},{"issue":"9657","key":"R7","doi-asserted-by":"publisher","first-page":"68","DOI":"10.1016\/S0140-6736(08)61706-7","article-title":"Burden and consequences of child maltreatment in high-income countries","volume":"373","author":"Gilbert","journal-title":"Lancet"},{"issue":"8","key":"R8","doi-asserted-by":"publisher","first-page":"637","DOI":"10.1001\/jama.1997.03550080047038","article-title":"Long-term effects of home visitation on maternal life course and child abuse and neglect. Fifteen-year follow-up of a randomized trial","volume":"278","author":"Olds","journal-title":"JAMA"},{"key":"R9","unstructured":"Child protection systems strengthening. UNICEF. 2021. URL: https:\/\/www.unicef.org\/documents\/child-protection-systems-strengthening [Accessed 18-06-2026]"},{"issue":"10","key":"R10","doi-asserted-by":"publisher","first-page":"763","DOI":"10.22454\/FamMed.2018.778329","article-title":"PedHITSS: a screening tool to detect childhood abuse in clinical settings","volume":"50","author":"Shakil","journal-title":"Fam Med"},{"issue":"6","key":"R11","doi-asserted-by":"publisher","first-page":"733","DOI":"10.1016\/s0145-2134(00)00137-x","article-title":"Assessing the value of structured protocols for forensic interviews of alleged child abuse victims","volume":"24","author":"Orbach","journal-title":"Child Abuse Negl"},{"key":"R12","doi-asserted-by":"publisher","DOI":"10.1016\/j.childyouth.2024.107939","article-title":"\u201cSuffering in silence\u201d: how social workers in child welfare practice experience and manage burnout","volume":"166","author":"Fuseini","journal-title":"Child Youth Serv Rev"},{"key":"R13","unstructured":"Peterson E . \u201cAn ecosystem of dysfunction:\u201d West Virginia still has a child welfare worker shortage, and it\u2019s taking a toll on foster kids and families. Mountain State Spotlight. 2025. URL: https:\/\/mountainstatespotlight.org\/2025\/01\/30\/foster-care-worker-shortage-stresses-system\/ [Accessed 12-10-2025]"},{"key":"R14","unstructured":"Key child abuse statistics for 2023. Ministry of Health and Welfare Child Abuse Response Division. 2024. URL: https:\/\/www.mohw.go.kr\/board.es?mid=a10107010000&bid=0040&act=view&list_no=1484071 [Accessed 12-10-2025]"},{"key":"R15","unstructured":"[Interview \u2460] child abuse caseworkers who struggle to last even one year. Save the Children. 2021. URL: https:\/\/www.sc.or.kr\/news\/storyView.do?NO=70936 [Accessed 12-10-2025]"},{"key":"R16","unstructured":"Achiam J Adler S Agarwal S et al. GPT-4 Technical Report. arXiv. Preprint posted online on Mar 4, 2024. [doi: 10.48550\/arXiv.2303.08774]"},{"key":"R17","doi-asserted-by":"crossref","unstructured":"Kang L Zhao Z Hsu D Lee WS . On the empirical complexity of reasoning and planning in LLMs. Presented at: Findings of the Association for Computational Linguistics; Nov 12-16, 2024; Miami, Florida, USA. 2024.[doi: 10.18653\/v1\/2024.findings-emnlp.164]","DOI":"10.18653\/v1\/2024.findings-emnlp.164"},{"key":"R18","doi-asserted-by":"crossref","unstructured":"Wang X Yuan P Feng S et al. CogLM: tracking cognitive development of large language models. Presented at: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics; Apr 29 to May 4, 2025; Albuquerque, NM. 2025.[doi: 10.18653\/v1\/2025.naacl-long.4]","DOI":"10.18653\/v1\/2025.naacl-long.4"},{"key":"R19","doi-asserted-by":"crossref","unstructured":"Kim H Gim G Kim Y et al. SAAS: solving ability amplification strategy for enhanced mathematical reasoning in large language models. Presented at: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing; Nov 12-16, 2024; Miami, FL, USA. [doi: 10.18653\/v1\/2024.emnlp-industry.15]","DOI":"10.18653\/v1\/2024.emnlp-industry.15"},{"key":"R20","unstructured":"ChatGPT. OpenAI. URL: https:\/\/chatgpt.com [Accessed 12-10-2025]"},{"key":"R21","unstructured":"deepseek. deepseek. URL: https:\/\/www.deepseek.com\/ [Accessed 12-10-2025]"},{"key":"R22","unstructured":"Google. Gemini. URL: https:\/\/gemini.google.com\/app [Accessed 12-10-2025]"},{"key":"R23","doi-asserted-by":"crossref","unstructured":"Steenstra I Nouraei F Arjmand M Bickmore T . Virtual agents for alcohol use counseling: exploring LLM-powered motivational interviewing. Presented at: IVA \u201924: Proceedings of the 24th ACM International Conference on Intelligent Virtual Agents; Sep 16-19, 2024; Glasgow, United Kingdom. Sep 16, 2024.[doi: 10.1145\/3652988.3673932]","DOI":"10.1145\/3652988.3673932"},{"key":"R24","doi-asserted-by":"publisher","DOI":"10.1145\/3712299","article-title":"LLM-based conversational AI therapist for daily functioning screening and psychotherapeutic intervention via everyday smart devices","author":"Nie","journal-title":"ACM Trans Comput Healthcare"},{"key":"R25","unstructured":"Izumi K Tanaka H Shidara K et al. Response generation for cognitive behavioral therapy with large language models: comparative study with socratic questioning. arXiv. Preprint posted online on Jan 29, 2024. [doi: 10.48550\/arXiv.2401.15966]"},{"key":"R26","unstructured":"KAVA 10th Anniversary: we remain steadfastly dedicatedto a world free from violence and abuse. Korean Association Against Violence and Abuse. URL: https:\/\/www.kava.kr\/joanne [Accessed 12-10-2025]"},{"issue":"8","key":"R27","doi-asserted-by":"publisher","first-page":"1453","DOI":"10.1176\/appi.ajp.160.8.1453","article-title":"Relationship between multiple forms of childhood maltreatment and adult mental health in community respondents: results from the adverse childhood experiences study","volume":"160","author":"Edwards","journal-title":"Am J Psychiatry"},{"key":"R28","unstructured":"Finkelhor D Turner HA Hamby SL Ormrod RK . Polyvictimization: children\u2019s exposure to multiple types of violence, crime, and abuse. Office of Juvenile Justice and Delinquency Prevention. URL: https:\/\/ojjdp.ojp.gov\/library\/publications\/polyvictimization-childrens-exposure-multiple-types-violence-crime-and-abuse [Accessed 21-05-2026]"},{"key":"R29","doi-asserted-by":"crossref","unstructured":"Gehman S Gururangan S Sap M Choi Y Smith NA . RealToxicityPrompts: evaluating neural toxic degeneration in language models. Presented at: Findings of the Association for Computational Linguistics: EMNLP 2020; Nov 16-20, 2020. [doi: 10.18653\/v1\/2020.findings-emnlp.301]","DOI":"10.18653\/v1\/2020.findings-emnlp.301"},{"key":"R30","doi-asserted-by":"crossref","unstructured":"Baheti A Sap M Ritter A Riedl M . Just say no: analyzing the stance of neural dialogue generation in offensive contexts. Presented at: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing; Nov 7-11, 2021; Online and Punta Cana, Dominican Republic. [doi: 10.18653\/v1\/2021.emnlp-main.397]","DOI":"10.18653\/v1\/2021.emnlp-main.397"},{"issue":"1","key":"R31","doi-asserted-by":"publisher","first-page":"295","DOI":"10.23047\/korea-sire.2025.04.295","article-title":"Developing child counseling data for AI learning and counseling tools for screening children at risk","volume":"5","author":"Kim","journal-title":"J Int Relat Interdiscip Educ"},{"key":"R32","unstructured":"AI-Hub. AI-Hub. URL: https:\/\/www.aihub.or.kr\/ [Accessed 12-10-2025]"},{"key":"R33","unstructured":"Korea National Information Society Agency. NIA. URL: https:\/\/www.nia.or.kr\/site\/nia_kor\/main.do [Accessed 12-10-2025]"},{"issue":"2","key":"R34","doi-asserted-by":"publisher","DOI":"10.2196\/mental.7785","article-title":"Delivering cognitive behavior therapy to young adults with symptoms of depression and anxiety using a fully automated conversational agent (Woebot): a randomized controlled trial","volume":"4","author":"Fitzpatrick","journal-title":"JMIR Ment Health"},{"issue":"4","key":"R35","doi-asserted-by":"publisher","DOI":"10.2196\/mental.9782","article-title":"Using psychological artificial intelligence (Tess) to relieve symptoms of depression and anxiety: randomized controlled trial","volume":"5","author":"Fulmer","journal-title":"JMIR Ment Health"},{"key":"R36","doi-asserted-by":"crossref","unstructured":"Malhotra G Waheed A Srivastava A Akhtar MS Chakraborty T . Speaker and time-aware joint contextual learning for dialogue-act classification in counselling conversations. Presented at: WSDM \u201922: Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining; Feb 21-25, 2022. [doi: 10.1145\/3488560.3498509]","DOI":"10.1145\/3488560.3498509"},{"key":"R37","doi-asserted-by":"crossref","unstructured":"Tavabi L Tran T Stefanov K et al. Analysis of Behavior Classification in Motivational Interviewing. In: Scherer S , editor. Presented at: Proceedings of the Seventh Workshop on Computational Linguistics and Clinical Psychology; Online. 2021.[doi: 10.18653\/v1\/2021.clpsych-1.13]","DOI":"10.18653\/v1\/2021.clpsych-1.13"},{"issue":"1","key":"R38","doi-asserted-by":"publisher","DOI":"10.2196\/59479","article-title":"The opportunities and risks of large language models in mental health","volume":"11","author":"Lawrence","journal-title":"JMIR Ment Health"},{"key":"R39","unstructured":"Liu JM Li D Cao H Ren T Liao Z Wu J . Chatcounselor: a large language models for mental health support. arXiv. Preprint posted online on Sep 27, 2023. [doi: 10.48550\/arXiv.2309.15461]"},{"key":"R40","doi-asserted-by":"crossref","unstructured":"Han H Park B Seo K . A self-determination theory-based career counseling chatbot: motivational interactions to address career decision-making difficulties and enhance engagement. Presented at: CHI EA \u201925: Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems; Apr 26 to May 1, 2025; Yokohama, Japan. [doi: 10.1145\/3706599.3720286]","DOI":"10.1145\/3706599.3720286"},{"issue":"3","key":"R41","doi-asserted-by":"publisher","DOI":"10.3390\/children12030359","article-title":"Conversational AI in pediatric mental health: a narrative review","volume":"12","author":"Mansoor","journal-title":"Children (Basel)"},{"key":"R42","doi-asserted-by":"crossref","unstructured":"Seo W Yang C Kim YH . ChaCha: leveraging large language models to prompt children to share their emotions about personal events. Presented at: CHI \u201924: Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems; May 11-16, 2024; Honolulu, HI, USA. [doi: 10.1145\/3613904.3642152]","DOI":"10.1145\/3613904.3642152"},{"issue":"10","key":"R43","doi-asserted-by":"publisher","DOI":"10.3390\/children10101659","article-title":"Artificial intelligence and child abuse and neglect: a systematic review","volume":"10","author":"Lupariello","journal-title":"Children (Basel)"},{"issue":"6","key":"R44","doi-asserted-by":"publisher","first-page":"1095","DOI":"10.1007\/s00247-022-05287-w","article-title":"Machine learning to identify distal tibial classic metaphyseal lesions of infant abuse: a pilot study","volume":"52","author":"Tsai","journal-title":"Pediatr Radiol"},{"key":"R45","doi-asserted-by":"publisher","DOI":"10.1016\/j.chiabu.2020.104755","article-title":"Can artificial intelligence achieve human-level performance? A pilot study of childhood sexual abuse detection in self-figure drawings","volume":"109","author":"Kissos","journal-title":"Child Abuse Negl"},{"issue":"2","key":"R46","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0247404","article-title":"A natural language processing and deep learning approach to identify child abuse from pediatric electronic medical records","volume":"16","author":"Annapragada","journal-title":"PLoS One"},{"issue":"12","key":"R47","doi-asserted-by":"publisher","first-page":"2326","DOI":"10.1016\/j.jpedsurg.2021.03.007","article-title":"Using deep learning and natural language processing models to detect child physical abuse","volume":"56","author":"Shahi","journal-title":"J Pediatr Surg"},{"issue":"6","key":"R48","doi-asserted-by":"publisher","first-page":"981","DOI":"10.1016\/j.acap.2021.11.004","article-title":"Development and validation of a natural language processing tool to identify injuries in infants associated with abuse","volume":"22","author":"Tiyyagura","journal-title":"Acad Pediatr"},{"issue":"1","key":"R49","doi-asserted-by":"publisher","first-page":"102","DOI":"10.1037\/1082-989x.5.1.102","article-title":"Neural network modeling of risk assessment in child protective services","volume":"5","author":"Marshall","journal-title":"Psychol Methods"},{"key":"R50","unstructured":"Data usage policy. AI-Hub. URL: https:\/\/aihub.or.kr\/intrcn\/guid\/usagepolicy.do?currMenu=151&topMenu=105 [Accessed 17-06-2026]"},{"key":"R51","unstructured":"Jhgan\/ko-sroberta-multitask. HuggingFace. 2020. URL: https:\/\/huggingface.co\/jhgan\/ko-sroberta-multitask [Accessed 17-05-2026]"},{"issue":"11","key":"R52","doi-asserted-by":"publisher","first-page":"205","DOI":"10.21105\/joss.00205","article-title":"hdbscan: hierarchical density based clustering","volume":"2","author":"McInnes","journal-title":"J Open Source Softw"},{"key":"R53","unstructured":"Hurst A Lerer A Goucher AP et al. GPT-4o system card. arXiv. Preprint posted online on Oct 25, 2024. [doi: 10.48550\/arXiv.2410.21276]"},{"key":"R54","unstructured":"Wei J Bosma M Zhao V et al. Finetuned language models are zero-shot learners. Presented at: Proceedings of the International Conference on Learning Representations; Apr 25-29, 2022; Virtual event. URL: https:\/\/openreview.net\/forum?id=gEZrGCozdqR [Accessed 18-06-2026]"},{"key":"R55","unstructured":"Tu W Deng W Zheng L Gedeon T . What does Softmax probability tell us about classifiers ranking across diverse test conditions?. arXiv. Preprint posted online on Jun 14, 2024. [doi: 10.48550\/arXiv.2406.09908]"},{"issue":"4","key":"R56","doi-asserted-by":"publisher","DOI":"10.3390\/e25040573","article-title":"Information entropy measures for evaluation of reliability of deep neural network results","volume":"25","author":"Gireesh","journal-title":"Entropy (Basel)"},{"issue":"1","key":"R57","doi-asserted-by":"publisher","first-page":"3","DOI":"10.2466\/pr0.1966.19.1.3","article-title":"The intraclass correlation coefficient as a measure of reliability","volume":"19","author":"Bartko","journal-title":"Psychol Rep"},{"key":"R58","unstructured":"evaluate-metric\/exact_match. HuggingFace. URL: https:\/\/huggingface.co\/spaces\/evaluate-metric\/exact_match [Accessed 19-05-2026]"},{"key":"R59","unstructured":"Geifman Y Uziel G El-Yaniv R . Bias-reduced uncertainty estimation for deep neural classifiers. Presented at: Proceedings of the 7th International Conference on Learning Representations; May 6 to Jun 9, 2019; New Orleans, Louisiana, USA. URL: https:\/\/openreview.net\/forum?id=SJfb5jCqKm [Accessed 18-06-2026]"},{"key":"R60","unstructured":"Paszke A Gross S Massa F Lerer A Bradbury J Chanan G et al. Pytorch: an imperative style, high-performance deep learning library. Presented at: Proceedings of the 33rd International Conference on Neural Information Processing (NeurIPS 2019); Dec 8-14, 2019; Vancouver, British Columbia, Canada. URL: https:\/\/proceedings.neurips.cc\/paper\/2019\/hash\/bdbca288fee7f92f2bfa9f7012727740-Abstract.html [Accessed 17-06-2026]"},{"key":"R61","unstructured":"The AI community building the future. HuggingFace. URL: https:\/\/huggingface.co\/ [Accessed 12-10-2025]"},{"issue":"1","key":"R62","doi-asserted-by":"publisher","first-page":"147","DOI":"10.1016\/S0364-0213(85)80012-4","article-title":"A learning algorithm for boltzmann machines","volume":"9","author":"Ackley","journal-title":"Cogn Sci"},{"key":"R63","unstructured":"LGAI-EXAONE\/EXAONE-35-78B-instruct. HuggingFace. 2024. URL: https:\/\/huggingface.co\/LGAI-EXAONE\/EXAONE-3.5-7.8B-Instruct [Accessed 12-10-2025]"},{"key":"R64","unstructured":"LGAI-EXAONE\/EXAONE-35-32B-instruct-AWQ huggingface.co. HuggingFace. 2024. URL: https:\/\/huggingface.co\/LGAI-EXAONE\/EXAONE-3.5-32B-Instruct-AWQ [Accessed 12-10-2025]"},{"key":"R65","unstructured":"An S Bae K Choi E Choi K Jungkyu Choi S Hong S et al. EXAONE 3.5: series of large language models for real-world use cases. arXiv. Preprint posted online on Dec 6, 2024. [doi: 10.48550\/arXiv.2412.04862]"},{"key":"R66","unstructured":"Monologg\/kobert. HuggingFace. 2024. URL: https:\/\/huggingface.co\/monologg\/kobert [Accessed 12-10-2025]"},{"key":"R67","unstructured":"Klue\/bert-base. HuggingFace. 2021. URL: https:\/\/huggingface.co\/klue\/bert-base [Accessed 12-10-2025]"},{"key":"R68","unstructured":"Klue\/roberta-base. HuggingFace. 2021. URL: https:\/\/huggingface.co\/klue\/roberta-base [Accessed 12-10-2025]"},{"key":"R69","unstructured":"Klue\/roberta-large. HuggingFace. 2021. URL: https:\/\/huggingface.co\/klue\/roberta-large [Accessed 12-10-2025]"},{"key":"R70","unstructured":"Monologg\/kobigbird-bert-base. HuggingFace. 2023. URL: https:\/\/huggingface.co\/monologg\/kobigbird-bert-base [Accessed 12-10-2025]"},{"key":"R71","unstructured":"vaiv\/kobigbird-roberta-large. HuggingFace. 2023. URL: https:\/\/huggingface.co\/vaiv\/kobigbird-roberta-large [Accessed 12-10-2025]"},{"key":"R72","unstructured":"BM-K\/KoSimCSE-roberta. 2022. URL: https:\/\/huggingface.co\/BM-K\/KoSimCSE-roberta [Accessed 12-10-2025]"},{"key":"R73","unstructured":"Qwen\/qwen25-3B-instruct. HuggingFace. 2024. URL: https:\/\/huggingface.co\/Qwen\/Qwen2.5-3B-Instruct [Accessed 12-10-2025]"},{"key":"R74","unstructured":"EleutherAI\/polyglot-ko-58b. HuggingFace. 2023. URL: https:\/\/huggingface.co\/EleutherAI\/polyglot-ko-5.8b [Accessed 12-10-2025]"},{"issue":"2","key":"R75","doi-asserted-by":"publisher","first-page":"153","DOI":"10.1007\/BF02295996","article-title":"Note on the sampling error of the difference between correlated proportions or percentages","volume":"12","author":"McNEMAR","journal-title":"Psychometrika"},{"issue":"2","key":"R76","doi-asserted-by":"publisher","first-page":"155","DOI":"10.1016\/j.jcm.2016.02.012","article-title":"A guideline of selecting and reporting intraclass correlation coefficients for reliability research","volume":"15","author":"Koo","journal-title":"J Chiropr Med"},{"key":"R77","unstructured":"Guo C Pleiss G Sun Y Weinberger KQ . On calibration of modern neural networks. Presented at: Proceedings of the 34th International conference on machine learning; Aug 6-11, 2017:1321-1330; Sydney, NSW, Australia. 2017.URL: https:\/\/dl.acm.org\/doi\/abs\/10.5555\/3305381.3305518 [Accessed 17-06-2026]"},{"issue":"1","key":"R78","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1175\/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2","article-title":"Verification of forecasts expressed in terms of probability","volume":"78","author":"Brier","journal-title":"Mon Wea Rev"},{"issue":"11-12","key":"R79","doi-asserted-by":"publisher","first-page":"1201","DOI":"10.1016\/j.chiabu.2007.03.021","article-title":"A structured forensic interview protocol improves the quality and informativeness of investigative interviews with children: a review of research using the NICHD Investigative Interview Protocol","volume":"31","author":"Lamb","journal-title":"Child Abuse Negl"},{"key":"R80","unstructured":"Child and adolescent counseling data. AI-Hub. URL: https:\/\/aihub.or.kr\/aihubdata\/data\/view.do?dataSetSn=71680 [Accessed 20-06-2026]"},{"key":"R81","unstructured":"Cuk-nlp\/conversational-AI-for-child-abuse-detection. GitHub. URL: https:\/\/github.com\/cuk-nlp\/Conversational-AI-for-Child-Abuse-Detection [Accessed 17-06-2026]"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2026\/1\/e86536","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,31]],"date-time":"2026-07-31T22:25:29Z","timestamp":1785536729000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2026\/1\/e86536"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,10]]},"references-count":81,"URL":"https:\/\/doi.org\/10.2196\/86536","relation":{},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,10]]},"article-number":"v28i4e86536"}}