S. Korea to build national AI library from master craftsmen's know-how, invest 20 trillion won in manufacturing AI

The government plans to convert the hands-on expertise of veteran manufacturing masters nearing retirement into data for developing AI models. The initiative aims to capture undocumented knowledge — from a skilled worker's fingertip instincts to the judgment calls made on the factory floor — in a national data library that will serve as a dedicated AI brain for the manufacturing sector.

The National AI Strategy Committee and related ministries unveiled the "Path to Korea's Manufacturing Transformation: Manufacturing AI 2030 Strategy" on Sunday, on the sidelines of a public briefing on three national mega-projects. The government set a target of investing 20 trillion won ($13 billion) in a public-private partnership by 2030 to generate more than 100 trillion won in economic value.

Securing manufacturing data is the strategy's starting point. AI transformation on the factory floor requires the simultaneous accumulation of process-specific data and on-site knowledge, yet data on equipment, processes and quality is scattered across individual companies at most domestic manufacturers, and small and medium-sized enterprises lack the personnel to refine and use it.

The government will collect high-quality manufacturing data and consolidate it in a "Manufacturing AX Data Library." The library will hold not only production process data but also the "tacit knowledge" that skilled workers have built up on the job — experience in detecting early signs of equipment failure, work sequences that reduce defect rates, and fine-tuning criteria for each process stage. The goal is to convert this undocumented know-how into data that AI systems can learn from.

Because manufacturing data contains a company's production conditions, quality-control methods and equipment operating information, sharing it between firms or allowing it to leak externally poses a significant risk of eroding its asset value. The government plans to design standardization, encryption and de-identification systems to establish a national-level management framework.

The collected data will feed the development of AI models tailored to the manufacturing sector. The government will build these models in stages — from lightweight models applicable to specific processes all the way to general-purpose foundation models covering the entire industry. Officials say AI models that understand the language of the factory floor and the flow of production processes are a prerequisite for broad adoption across key industries including semiconductors, shipbuilding and automobiles.

At an advanced stage, the government will pursue development of "full-stack AI factory" technology — large AI agents capable of controlling the full range of manufacturing production and management activities, including factory and product design, verification and trial production, factory operations, and retail and logistics. The plan also calls for expanding development of key components and pilot projects aimed at commercializing manufacturing-specific humanoid robots.

Securing foundational technology for manufacturing physical AI is also on the agenda. Target areas include AI models linked to the laws of physics, device-to-robot connectivity, low-latency and high-reliability communication networks, and integrated security technology spanning the full lifecycle of models, AI agents and control systems. Officials noted that for AI to make judgments and control equipment and robots inside a factory, it requires a different kind of technology stack from general generative AI.

Industrial complexes will serve as the primary hubs for rolling out the strategy. The government will establish "M.AX clusters" to transform industrial complexes — which account for two-thirds of domestic manufacturing output and exports, and half of manufacturing employment — into AI-enabled industrial zones. Plans include building shared infrastructure such as pilot testbeds and edge computing centers at each complex, as well as promoting AI smart factory projects built on cooperation between large companies and their suppliers.

Whether small and medium-sized enterprises participate will determine the strategy's success. According to the Ministry of SMEs and Startups, SMEs make up 99.6 percent of all manufacturing companies. Recognizing that smaller manufacturers cannot easily secure AI talent and infrastructure on their own, the government will run parallel programs covering large-SME cooperation, collaboration among SMEs themselves, and on-site worker education.

An investment and certification framework to build the manufacturing AI ecosystem will also be put in place. The government will use funds and guarantees to encourage private investment and will link the initiative to the 150 trillion won National Growth Fund. A "Manufacturing AX Certification" system will be introduced to foster specialized manufacturing AI companies. On the legal side, the government will pursue amendments to the Industrial Digital Transformation and AI Utilization Promotion Act and the SME Smart Manufacturing Innovation Act.

Deputy Prime Minister and Minister of Science and ICT Bae Kyung-hoon said manufacturing AI "must advance to a stage where AI understands the physical phenomena and process flows of the factory floor and can independently judge and control equipment and robots," adding that the government would "actively support the development of manufacturing AI into technology that actually works on real factory floors."

Minister of Trade, Industry and Energy Kim Jung-kwan said "M.AX — the AI transformation of manufacturing — is a core challenge that will determine the future competitiveness and survival of our manufacturing sector." He stressed that "since this plan was developed through six months of deliberation by related ministries and private-sector experts, it is crucial that each ministry implement its tasks responsibly so that real value is created in the real world."

Vice Minister of SMEs and Startups Noh Yong-seok said the AI transformation of small and medium-sized enterprises, which account for 99.6 percent of all manufacturing companies, "is an essential condition for the AI transformation of Korea's manufacturing sector as a whole." He pledged to "build a support system so that specialized manufacturing AI companies can grow from startup and investment all the way to global expansion."