OpenAI is hosting the first hackathon in Korea to solve practical problems in the social welfare field using AI tools. This is an opportunity to verify whether social welfare practitioners who are not developers can create actual working prototypes using ChatGPT Work and Codex.

OpenAI announced on the 24th that it will hold a mini hackathon on the 25th at the Woosuk Economics Building of Seoul National University in collaboration with the Hyundai Motor Chung Mong-Koo Foundation and the Social Innovation Impact Lab (SNU CSI) of the Seoul National University Institute of Social Welfare.

This event was organized as part of the second core competency training of the 'CMK Social Welfare Innovation Leader Academy,' hosted by the Hyundai Motor Chung Mong-Koo Foundation and operated by SNU CSI.

A total of 30 people, including social welfare field practitioners and undergraduate and graduate students majoring in related fields, will participate in five teams.

Participants will first structure the field problem and write a Product Requirements Document (PRD), and then implement a prototype in the form of a webpage or app using the ChatGPT work and the agentic capabilities of Codex.

Development topics by team include: a tool to support visiting care records and handover; a mental health recovery-oriented case record review tool; an app to identify the living relationship networks of single-person households; a tool to track changes by session for educational welfare recipients; an errand sharing platform for ward guardians; and an app for early detection of daily risks in discharged elderly patients.

OpenAI Korea engineers provide on-site coaching throughout the entire prototype implementation process, including setting the development scope, diagnosing the cause of errors, improving code and prompts, and checking accessibility and safety.

Ko Ki-seok, Head of Policy at OpenAI Korea, stated, “The value of AI comes not from the technology itself, but from helping people better define and solve real-world problems,” adding, “This program enables social welfare field practitioners without development experience to utilize ChatGPT Work and Codex to turn ideas from the field into working prototypes.” "It shows that it can be developed," he said.
The hackathon results are planned to be further developed through team-based action learning projects from August to October.

Participants plan to develop the solution by jointly reviewing field applicability and limitations, as well as privacy and ethical considerations.