"The gap between companies with access to data, GPUs, specialized personnel, and on-site demonstration environments, and those without, could ultimately lead to a gap in manufacturing competitiveness."
This was stated by Byeng Dong - Youn, CEO of onepredict and professor of Mechanical Engineering at Seoul National University, as he pointed out the "manufacturing AX (AI Transformation) polarization" facing the domestic manufacturing industry. He spoke as a presenter at the 'Policy Debate for Manufacturing AI Transformation and On-site Expansion' held at the National Assembly Member's Office Building on the 23rd. He argued that for AI to go beyond being the exclusive property of a few leading companies and permeate into small and medium-sized manufacturing sites, the focus of policy must shift from R&D and one-time demonstrations to 'public infrastructure' and 'actual on-site operations.'
Co-hosted by Representative Kim Won-i of the Democratic Party of Korea and Representative Choi Hyung-du of the People Power Party, the debate was organized for industry, academia, and research experts to gather and discuss ways to expand manufacturing AI in the field. CEO Youn, presenting on behalf of the industry, put forward policy proposals based on his experience with AI transformation in manufacturing sites.
He pointed out, "Even if data is piled up on-site, it is often scattered by equipment or in a format that AI cannot learn from." To resolve this, he emphasized the need for educational support, public support, and the provision of infrastructure. In particular, he noted that the infrastructure to actually implement AX is the most crucial element. This is because relatively small and medium-sized enterprises (SMEs) often lack the infrastructure even if they wish to adopt AX.
"The industrial infrastructure of the AI era is the GPU," CEO Youn said. "We must provide support so that high-performance computing resources, which are difficult for SMEs to build individually, can be utilized jointly at industrial complexes or regional hubs."
CEO Youn also stressed the importance of turning data into assets. He explained, "99.9% of on-site data is consumable data used only for control purposes and is never opened again. South Korea can secure the world's largest manufacturing data territory only when we capitalize and recycle this data through a data grading system."
He argued that three types of support must be provided to resolve the AX polarization: △Retraining for current employees, △Construction of a joint GPU center and support for voucher programs, and △Digitalization of highly skilled workers' know-how (tacit knowledge) and the provision of demonstration testbeds.
Kim Taek-soo, acting president of the Korea Institute of Industrial Technology, who attended the debate, said, "When you visit SME sites, there are many cases where they don't even know what data is or how to acquire it. We must provide customized support by level to help them lay the foundation to embrace future manufacturing technologies."
...
...
...
...
This was stated by Byeng Dong - Youn, CEO of onepredict and professor of Mechanical Engineering at Seoul National University, as he pointed out the "manufacturing AX (AI Transformation) polarization" facing the domestic manufacturing industry. He spoke as a presenter at the 'Policy Debate for Manufacturing AI Transformation and On-site Expansion' held at the National Assembly Member's Office Building on the 23rd. He argued that for AI to go beyond being the exclusive property of a few leading companies and permeate into small and medium-sized manufacturing sites, the focus of policy must shift from R&D and one-time demonstrations to 'public infrastructure' and 'actual on-site operations.'
Co-hosted by Representative Kim Won-i of the Democratic Party of Korea and Representative Choi Hyung-du of the People Power Party, the debate was organized for industry, academia, and research experts to gather and discuss ways to expand manufacturing AI in the field. CEO Youn, presenting on behalf of the industry, put forward policy proposals based on his experience with AI transformation in manufacturing sites.
He pointed out, "Even if data is piled up on-site, it is often scattered by equipment or in a format that AI cannot learn from." To resolve this, he emphasized the need for educational support, public support, and the provision of infrastructure. In particular, he noted that the infrastructure to actually implement AX is the most crucial element. This is because relatively small and medium-sized enterprises (SMEs) often lack the infrastructure even if they wish to adopt AX.
"The industrial infrastructure of the AI era is the GPU," CEO Youn said. "We must provide support so that high-performance computing resources, which are difficult for SMEs to build individually, can be utilized jointly at industrial complexes or regional hubs."
CEO Youn also stressed the importance of turning data into assets. He explained, "99.9% of on-site data is consumable data used only for control purposes and is never opened again. South Korea can secure the world's largest manufacturing data territory only when we capitalize and recycle this data through a data grading system."
He argued that three types of support must be provided to resolve the AX polarization: △Retraining for current employees, △Construction of a joint GPU center and support for voucher programs, and △Digitalization of highly skilled workers' know-how (tacit knowledge) and the provision of demonstration testbeds.
Kim Taek-soo, acting president of the Korea Institute of Industrial Technology, who attended the debate, said, "When you visit SME sites, there are many cases where they don't even know what data is or how to acquire it. We must provide customized support by level to help them lay the foundation to embrace future manufacturing technologies."