Generative AI Chatbot for Public Information and Web Content Guidance

Overview

Through an analysis of the official and sub-websites of the Korea Institute for Gender Equality Promotion and Education (KIGEPE) and its customer consultation operations, this service provides customized gender equality content.

Public demand for information on preventing the "Five Major Types of Violence" (power-based sexual violence, digital sexual violence, domestic violence, dating violence, and stalking) is increasing annually. Consequently, it was necessary to develop a system that integrates prevention education and victim support to provide real-time information to the public.

Project description

The AI Chatbot Service provides the following features:

• Content Search: Prevention education, violence prevention, and crime types
• Civil Complaint Filing: Reporting damage and victim support
• Service Requests: Modification or deletion of information and data -History Management: Customized support based on past user history

The project involved crawling notices and educational announcements from the official website and its sub-sites (Education Center, E-learning Center, Gender-On, and D-Clean). By training on a diverse dataset of video, image, and text-based educational content, this service was configured as a chatbot that provides guidance on KIGEPE (Korea Institute for Gender Equality Promotion and Education) and gender equality training, enabling users to locate specific content immediately.

Key technological features

• Data Indexing Optimization: Focused on enhancing the relevance and accuracy of results through optimized data indexing and embedding
• RAG Utilization: Optimized the retrieval phase, flexibly applied various query strategies, and utilized embedding similarity
• Implementation of Advanced RAG: To overcome the limitations of Naive RAG (which simply combines retrieval and generation), Advanced RAG technology was applied to implement sophisticated search functions and context-based response generation.

Expected outcomes

This AI chatbot service has streamlined consultation procedures through automated data collection and analysis. It is expected to significantly reduce the overall workload required for handling consultations and inquiries.

Launch year: 2024

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.2, https://eng.nia.or.kr/site/nia_eng/ex/bbs/View.do?cbIdx=31975&bcIdx=29182&parentSeq=29192