At the "AI Champion Hackathon" held at KINTEX in Ilsan, Goyang, on the 23rd–24th of last month, the Korea International Cooperation Agency (KOICA) team ranked side by side. Among the 24 teams representing government ministries, local governments, and public institutions, eight advanced to the finals, and two of them were from KOICA.
KOICA's "Four Sons Two Daughters" team won the grand prize, equivalent to runner-up, and "Urgent Now" won the excellence prize, equivalent to third place. All four participants on the two teams are members of "KoSam," KOICA's in-house coding club. The name means studying coding like a third-year high school student studies.
The final assignment was to create a solution to help regions facing population extinction through "vibe coding," in which AI writes code based on natural language commands. The "Four Sons Two Daughters" team developed a location-based program called "Appago," which helps families of transferred public institution employees adapt to the region, in four hours. "Urgent Now" unveiled a solution that provides education and transportation information to Gwangju and South Jeolla residents to help them choose a residence.
How did KOICA employees, who carry out official development assistance (ODA) projects for developing countries, learn AI and coding? On the 13th at KOICA headquarters in Pangyo, Seongnam, we met Directors Jeon Damin and Kim Ban-seok of the "Four Sons Two Daughters" team; Director Kim Jong-jun and Assistant Manager Heo Se-hyeon of "Urgent Now"; and Assistant Manager Kim Woo-young of KOICA's Artificial Intelligence Transformation Office data team. The following is a Q&A.
─Both teams performed well at the hackathon. What was the secret?
Kim Woo-young said, "Starting in 2024, KOICA conducted AX (artificial intelligence transformation) capacity-building training for all employees. Rather than focusing on short-term results, we set strategy from a long-term perspective and moved to cultivate talent. We also supported an in-house coding club so employees could try various technologies in a free environment.
"Through a program for proposing work improvements, we had employees develop automation solutions that solve administrative inconveniences with coding and computer technology. We provided incentives for outstanding proposals. I think these efforts accumulated to strengthen AI utilization capacity within the organization."
─Please introduce the in-house coding club "KoSam."
Jeon Damin said, "We have about 10 members. The four who competed in this hackathon are all part of KoSam. We named it 'KoSam' to mean we study coding hard like third-year high school students do. Even if they are not former developers, interested employees gather to study coding and build programs needed for actual work together."
─Do you have an example of developing an actual program through club activities?
Jeon Damin said, "We developed 'All PDF,' which converts multiple Hangul documents into PDF files all at once, and shared it internally. Because of the nature of KOICA's work, we often place orders with outside vendors. We also created a program that lets you manage ordering schedules at a glance in a calendar format."
Kim Jong-jun said, "While working in the budget office, I directly developed programs needed for the job. I built a 'budget, accounting, and finance chatbot' that can search extensive budget, accounting, and finance regulations and transaction histories, and made it available on work PCs in a segmented network.
"I also created 'Smart Allocation,' which supports AI-assisted decision-making for budget allocation. KOICA employees travel frequently to developing countries, and the vaccines required before departure differ by country. Based on that, I developed 'Vaccine Pass,' which lets you immediately check essential and recommended vaccines by country."
─Can the government's ODA projects be carried out more efficiently by using AI?
Kim Jong-jun said, "If we systematically collect and analyze partner countries' data using AI, we can execute official aid more efficiently. It can be used to determine which regions and sectors urgently need aid and to measure outcomes after projects."
"I am personally scheduled to be dispatched to Tanzania soon. On the ground, data is still often written and managed by hand. Before introducing AI, we need to start by digitizing data that is created and managed manually."
─Demand for transferring AI technology likely is high in developing countries as well.
Kim Woo-young said, "The key is introducing AI at an appropriate level that considers the environment of developing countries. Bringing over the latest technology used in advanced countries does not mean it will work on the ground. That is because internet and power infrastructure, levels of data accumulation, and the availability of skilled personnel vary by country.
"In addition to classifying partner countries by income level, we need to closely analyze each country's technological and social environment. Rather than introducing high-performance systems all at once, we should transfer systems step by step, starting at levels that can actually be operated and maintained."
─What role can KOICA play in the process of transferring AI technology?
Kim Woo-young said, "We should not stop at simply providing AI programs or equipment. We need to co-design everything from how local officials and practitioners collect and manage data to how AI is applied to actual administration and public services.
"For ODA projects to be sustainable, partner countries must be able to operate them on their own after the project ends. Because KOICA understands each country's administrative systems and local needs, it can play a role in consolidation between technology and the field."
─Many in the public sector still find AI difficult. What advice would you offer?
Kim Jong-jun said, "We need to break free from vague fears about AI. The more someone fears that 'when the AI era comes, my value will fall,' the more that person should try it directly. Once you actually use it, you see what AI does well and what it does not. When you learn its limits, the vague fear also diminishes."
Heo Se-hyeon said, "Most people say AI is important, but not many use it frequently. Even when they do, it often stops at searching or summarizing documents. It seems the hardest part is taking the first step of installing a program once and running it yourself."
Kim Ban-seok said, "It is important to create opportunities for public officials and public institution employees who have not encountered AI to experience it naturally. I hope competitions and training like this hackathon, where even people without much development experience can use AI to implement ideas, are held often."