Development of a Sexual Crime Crisis Response (Investigation, Safety, Support) Platform

Overview

The Gwangju Metropolitan Police Agency anticipates over 30,000 crimes annually, yet public support provided through its victim protection and support unit (8 personnel) amounts to only about 900 cases per year, representing roughly 3% of the total. Due to insufficient police personnel (Gwangju Metropolitan City has a total of 5 police stations), delays occur in linking investigations, safety measures, and support. Furthermore, frequent personnel transfers lead to variations in the content of victim support provided, depending on the capabilities of individual officers.

For socially vulnerable groups such as children, adolescents, the elderly, and persons with disabilities, reporting sexual crimes is often delayed due to an inability to recognize the harm, or even among the general public, due to social stigma, shame, helplessness, post-traumatic stress, the burden of legal procedures, perpetrator intimidation, distance, or hospital admission. This delay raises concerns about additional harm during this period. Furthermore, victims often lack the specific awareness needed to search through the scattered online and offline information available.

A support system is therefore required that enables any sexual crime victim to receive specialized, tailored crisis intervention services—including investigation, safety, support—at any time, and from anywhere, independent of the capabilities of police officers or experts.

Project description

This project develops an 'AI-based crisis intervention (investigation/safety/support) service' that utilizes LLM to provide non-face-to-face and assessment of victim details, along with personalized 1:1 solution. These solution cover guidance on investigation procedures, economic, housing, legal, medical, psychological, and safety support, helping victims recognize the severity of their situation and support the prompt initiation of investigations.

Additionally, an 'AI-based administrative support service' for police officers and experts will be developed. This service will provide LLM-based summary reports (including investigation guides such as details of harm and classification of sex crime types, legal support including precedents, assessment of harm severity, and information on protection and support agencies) to support rapid crisis identification and enhanced expertise.

Key technological features

• Utilizing hyper-scale text data, pre-trained large language models (LLMs) are employed for various natural language processing tasks, including summarizing victim statements, classifying types of sex crimes, and providing support for similar precedents and legal information.

Expected outcomes

Development of this service is scheduled to begin in 2025. It is expected to enhance the government's crisis intervention capabilities for sex crime victims, thereby improving the level of safety support for national society.

The service will be provided in two formats: text and video. A text version will be available for use in open spaces such as public areas, while a video version will be provided for users who have difficulty with text input, including the elderly, children, individuals with intellectual disabilities, or the visually impaired, as well as those who prefer face-to-face service.

This will support preventing unnecessary investigations through pre-screening before complaint acceptance, allowing investigative resources to focus on crimes. It will also provide swift procedures for victims requiring police protection, ensuring their personal safety.

Launch year: 2025

Attribution

Case study courtesy of the National Information Society Agency, Republic of Korea, Apolitical's Content Partner. Sourced from Use Cases of Public AI Service, Vol.1,  https://eng.nia.or.kr/site/nia_eng/ex/bbs/View.do?cbIdx=31975&bcIdx=28997&parentSeq=29009