The South Korean government is advancing its manufacturing artificial intelligence (AI) transition, known as M.AX, with the establishment of AI factories that are showing improvements in productivity and defect rates. The government aims to build 500 AI factories by 2030, converting manufacturing data and skilled workers' expertise into AI models to promote M.AX.
On August 4, the Ministry of Trade, Industry and Energy announced key achievements and future plans for the M.AX strategy during a business report. Last September, the government launched the M.AX Alliance, involving over 1,000 organizations, including manufacturing companies, AI firms, universities, and research institutions.
In February, a new division called Sandan AX was established to promote regional AX initiatives centered around industrial complexes, creating a system with 11 divisions, including AI factories, AI semiconductors, and AI robotics. The number of participating companies and institutions has increased by 50% to over 1,500 within eight months of the launch.
Concrete results have been confirmed in the field. The Ministry supported AX initiatives in over 170 business sites through its AI factory pilot projects. Among these, 42 sites that began to show visible results reported an average productivity increase of 30.1% and a 15.5% reduction in defect rates.
Industry-specific examples have emerged. In the semiconductor sector, Shinhan Diamond implemented an AI automatic inspection system for CMP discs used in wafer polishing, reducing quality inspection time by 33%. Surplus Global integrated AI into the semiconductor equipment recycling process, boosting productivity by 66.7%.
In the automotive sector, Asan Seongwoo High-tech reduced quality inspection time by 90% through an autonomous manufacturing system for electric vehicle batteries, while Hwasin lowered product defect rates by 20% by introducing a predictive maintenance system in the battery case production process.
In shipbuilding, HD Hyundai Samho enhanced productivity by 50% by adopting an AI-based autonomous welding and inspection solution for rudder trunk welding and production processes. ENK utilized an AI vision inspection system to cut inspection time for high-pressure cylinders used in shipbuilding and special gases by 22%.
The Ministry is expanding the application of M.AX from individual manufacturing processes to products and industrial complexes. In the product AX sector, it has begun developing and testing industry-specific humanoids for hazardous or repetitive tasks, such as fire monitoring and extinguishing in shipbuilding workplaces, valve operations in toxic gas environments in the chemical industry, and product packaging in logistics.
The government also aims to support the establishment of 500 AI factories by 2030, focusing on high-quality manufacturing data collection and AI model development and testing. It plans to develop core technologies for full-stack AI factories that operate the entire manufacturing process autonomously without human intervention, with a goal of commercialization and export after domestic testing.
Efforts will also be made to convert the experience and expertise of skilled workers into datasets. The Ministry plans to develop and validate AI models and solutions that learn from the tacit knowledge of master craftsmen and skilled workers, and to establish a national manufacturing data library for storing, managing, and utilizing high-quality manufacturing data.
Minister of Trade, Industry and Energy Kim Jeong-kwan stated, "The results of increasing productivity and reducing defect rates have been confirmed in the field, and we will ensure that verified models are rapidly disseminated to small and medium-sized enterprises and regional industrial complexes. We will actively support the accumulation and utilization of manufacturing data and the expertise of skilled workers, and implement full-stack AI factories operated by AI and robots, enabling our country to leap forward as a global leader in M.AX by 2030."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.