The Korea Electronics Technology Institute (KETI) has launched a full-scale industry-academia-research cooperation system to transform the domestic manufacturing industry into an AI-based one, and has begun jointly developing a manufacturing-specific AI Foundation Model (MFM).

On the 14th, KETI signed a business agreement with Seoul National University, KAIST, Pohang University of Science and Technology (POSTECH), OnePredict Inc., and IniGee Inc. in Pangyo, Gyeonggi Province, for joint research on a manufacturing-specific AI foundation model. This collaboration is part of the "Development of Core Technologies for MFM-Based SDM Platforms for AI Factory Advancement," a project funded by the Ministry of Trade, Industry and Energy and the Korea Institute of Industrial Technology Evaluation and Planning (KIET) and projected to reach KRW 29.2 billion by 2028.

The Manufacturing-Specific AI Foundation Model (MFM) is an AI model that internalizes manufacturing domain knowledge by learning time-series data and machine vision-based image data generated in the manufacturing process. Unlike general-purpose language models (LLMs), it can implement high-reliability AI functions by learning from data generated in actual manufacturing sites.

KETI has formed a research team of approximately 100 researchers, centered around the Autonomous Manufacturing Research Center, in collaboration with leading universities in industrial AI, including Seoul National University, KAIST, and Pohang University of Science and Technology (POSTECH), as well as AI-focused autonomous manufacturing companies. They plan to support the manufacturing industry's AI transformation through joint development of MFM, sharing of data and verification infrastructure, technical support, and joint lab operations.

KETI is also developing a software-defined manufacturing (SDM) platform that enables flexible changes to manufacturing operations based on MFM. The platform is designed to directly create and automatically execute field-specific AI models through an AI agent and conversational interface. It also includes various core technologies, including digital twin-based process configuration, edge equipment data collection, and security incident prevention.

The developed MFM and SDM platforms will be applied to core processes in major domestic industries, including automobiles, oil refining/petrochemicals, and semiconductor equipment, and their performance and effectiveness will be verified through testbeds. In particular, model performance will be continuously improved using manufacturing data accumulated through the Ministry of Trade, Industry and Energy's AI Factory Pilot Project. Upon completion, the platforms will be provided to participating companies.

Meanwhile, KETI is building a 'Manufacturing AI Solution Development Support Center' within the Gyeonggi Enterprise Growth Center in Seongnam City. The center will provide ongoing support for manufacturing-specific AI model learning, solution verification, and corporate-tailored AI adoption. It will also operate a manufacturing AI developer training program, contributing to addressing the shortage of specialized personnel in industrial settings.

KETI Autonomous Manufacturing Research Center Director Song Byeong-hun stated, “The use of AI technology is essential to secure the future competitiveness of the manufacturing industry,” and added, “We will lead a new leap forward for the domestic manufacturing industry through the successful development and industrial expansion of a manufacturing-specific foundation model.”