AI Learns Skilled Workers' 'Fingertip Feel'... Hanyang University ERICA to Build a KRW 4.4 Billion Manufacturing AI Guide

[Financial News] Technology development is gaining momentum to address labor shortages in manufacturing by converting skilled workers' years of experience and 'fingertip feel' into artificial intelligence (AI) data.
A research team led by Professor Jaewon Lee of the Department of Battery Materials and Chemical Engineering at Hanyang University ERICA announced on the 30th that it had begun developing an 'Intelligent Mold Guidance System' with Jaikonsi, Aizen and others. The system will convert skilled workers' know-how into AI data and provide real-time guidance to workers on the shop floor. The research was selected for the Ministry of Trade, Industry and Energy's (MOTIE) AX-Sprint Program, with KRW 4.385 billion in government R&D funding to be invested through the end of 2027.
Mold manufacturing, which produces molds for parts used in automobiles and home appliances, relies heavily on polishing work at the final stage, when surfaces are smoothed and finished. Because the pressure and angle of tools, as well as the polishing direction, depend largely on skilled workers' accumulated judgment, product quality can vary from worker to worker, often requiring rework. In particular, the aging of the workforce and the declining influx of young workers have recently placed the transfer of these skills at risk.
The system being developed will have AI learn veterans' subtle movements and surface-processing data, then provide workers with real-time guidance on appropriate work methods through visual displays and other interfaces. Once introduced, the system is expected to reduce quality gaps caused by differences in worker skill levels and significantly shorten the time required for beginners to become experienced workers.
The Hanyang University ERICA research team is responsible for developing the foundational technology that enables AI to automatically analyze complex three-dimensional mold drawings and identify key areas requiring machining, such as flat and curved surfaces. Drawing analysis is the first step in determining the appropriate tools and processes and is central to the overall system.
The project will be carried out as an industry-academia-research consortium involving Jaikonsi, the lead company overseeing software integration; Korea Institute of Industrial Technology (KITECH); Korea University (KU); Texas A&M University in the United States; and demand companies Korea Mold and JT. The participating organizations will pursue technology development, validation at actual manufacturing sites and commercialization in parallel.
Professor Jaewon Lee of Hanyang University ERICA said, "This is a meaningful attempt to convert nonstandardized experience that had remained only in the minds and fingertips of skilled workers into a quantitative data asset through AI technology," adding, "We will build a manufacturing AI model that addresses labor shortages and gaps in skills transfer at mold manufacturing sites and can later be expanded to similar foundational manufacturing processes."  

[email protected] Man-ki Kim Reporter