World Job Plus AI Demonstration for Citizen Service and Operational Support

Overview

The burden of human and material resources required to support diverse languages for users of 'World Job Plus'—which provides integrated information on overseas employment, entrepreneurship, volunteering, and other international ventures—has increased.

Accordingly, the Korea Human Resources Development Service (KHRD) sought to provide ongoing overseas employment support services, including AI-based chatbot customer service, to job seekers using World Job Plus by leveraging super-large AI technology.

Project description

An AI service was developed for public customer service and operational support by utilizing key data from the Korea Employment Agency (World Job Plus inquiry board data, various announcement documents, job seeker-related information).

It was built and processed training data from collected information, then trained and constructed a language model to implement a resume editing GPT capable of handling inquiries and complaints while supporting multiple languages.

Additionally, a multilingual resume-based mock interview question generation GPT was implemented to enable 24/7 customer support without time or location constraints, allowing job seekers to access overseas employment support services.

Key technological features

• Collecting Q&A-format training data and performing data preprocessing using PDF annotation tools
• Implemented a retrieval engine and developed integration with the maumGPT baseline model and fine-tuned models
• Building a unified Korea Employment Agency-specialized GPT model through individual API packaging of the search engine and LLM model and creating an integrated workflow

Expected outcomes

World Job Plus is expected to significantly enhance the efficiency of public service delivery and drastically reduce the time required for the Korea Human Resources Development Service's public inquiry response operations by providing a GPT model for citizen-facing interactions and operational support.

Launch year: 2023

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=29019