AI Document Integration in Police Administration: Security Challenges and Policy Implications
Korean Association of Public Safety and Criminal Justice · 30.06.2026 · 35 (2) · 151-182
DOI: 10.21181/kjpc.2026.35.2.151 ↗Kurz gefasst
Fachartikel in der Zeitschrift Korean Association of Public Safety and Criminal Justice, online erschienen am 30.06.2026. Verfasst von Wongi Choi und 1 weiteren Autor:innen.
Abstract (Originalsprache)
This study examines the security risks and institutional requirements associated with adopting generative AI document tools in police organizations, with a particular focus on administrative affairs functions. The analysis is limited to document-intensive tasks such as planning memos, event preparation documents, meeting summaries, and the organization of rules and legal materials. The study uses literature review, analysis of public-sector AI adoption cases, review of official security and privacy guidelines, and a qualitative suitability assessment of document workflows. The findings indicate that AI document tools are highly useful for repetitive drafting, summarization, comparison of similar documents, and retrieval support based on approved materials. However, documents related to personnel, budget, audit, security, and investigative matters contain highly sensitive information, and inputting such content into external SaaS-based AI services may create serious risks, including data leakage, hallucination, unclear accountability, and automation bias. Accordingly, the study concludes that a controlled architecture based on a closed network or on-premise small language model combined with retrieval-augmented generation is more appropriate for police organizations than unrestricted use of open cloud services. The paper contributes by discussing police AI adoption not only in terms of efficiency but also through the integrated lens of security, governance, and institutional design.
Zitieren
APA
Choi, W., & Roh, S. (2026). AI Document Integration in Police Administration: Security Challenges and Policy Implications. Korean Association of Public Safety and Criminal Justice, 35(2), 151-182. https://doi.org/10.21181/kjpc.2026.35.2.151
Vancouver
Choi W, Roh S. AI Document Integration in Police Administration: Security Challenges and Policy Implications. Korean Association of Public Safety and Criminal Justice. 2026;35(2):151-182. doi:10.21181/kjpc.2026.35.2.151
BibTeX
@article{Choi2026,
author = {Choi, Wongi and Roh, Seungkook},
title = {AI Document Integration in Police Administration: Security Challenges and Policy Implications},
journal = {Korean Association of Public Safety and Criminal Justice},
year = {2026},
volume = {35},
number = {2},
pages = {151--182},
doi = {10.21181/kjpc.2026.35.2.151}
}
EUVIMED