Support for AI-Based Proactive Response System Against Online Sexual Exploitation of Children and Youth

Overview

Recently, digital sex crimes targeting children and adolescents have rapidly spread, primarily in online environments, increasing their social severity.

The number of teenagers victimized by digital sex crimes surged from 111 in 2018 to 2,209 in 2023, with the nature of these crimes rapidly evolving. In 2024, the use of deepfake technology to create false sexual content has increased significantly, with minors accounting for 48.8% of victims, reaching alarming levels. This project aims to establish a system that swiftly responds to digital sex crimes by leveraging artificial intelligence (AI) to analyze text and images in real-time online.

This system will use text mining and data mining technologies to identify warning signs, automatically block or report suspected sexual exploitation content, and prevent the spread of harm proactively.

Project description

The project aims to develop a system that proactively prevents harm by monitoring online luring and suspicious information related to child and adolescent sexual exploitation, then reporting and supporting its removal. AI can rapidly analyze large volumes of data that are difficult for humans to process manually and detect risk indicators in real time. Specifically, by utilizing text mining technology, it can identify attempts at online luring and grooming early on and structure meaningful patterns, enabling swift response.

Furthermore, through data mining-based analysis, it learns and detects not only existing patterns but also modified or newly emerging types of sexual exploitation content, enabling the automatic blocking or reporting of suspicious content. By operating online counseling channels for victims, it enables the early detection of children and adolescents at high risk of sexual exploitation and provides methods to prevent the spread and prolongation of sexual exploitation.

Key technological features

• Image Analysis (OCR): Applying image-based character recognition technology (OCR) to identify text within online sexual exploitation content and developing models adaptable to diverse image environments
• Self-Learning Text Classification Technology: Utilizes TA Analytics (content classification, text mining, natural language processing) technology to automatically classify collected text and detect suspected sexual exploitation expressions in advanced forms through iterative learning

Expected outcomes

Starting in 2025, the project plans to establish a self-learning online sexual exploitation proactive response artificial intelligence (AI) system. The system will automatically collect online content and create an environment for continuous, automated monitoring of sexual exploitation materials 24 hours a day, 365 days a year. Furthermore, the service aims to prevent the spread of harm through the swift detection and blocking of illegal and harmful content, enhance victim protection, and thereby improve public trust.

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=29008