Gov't-led AI manufacturing push needs company-first approach, research body says

A government-led push to boost the competitiveness of South Korea's manufacturing sector through AI transformation needs to be tailored to the actual needs of businesses — and must lower barriers so small and medium-sized enterprises, not just large conglomerates, can participate broadly, a research body has said.

The Korea Institute for Advancement of Technology, known as KIAT, made the assessment in a recently published report on manufacturing data-based AI transformation trends and policy implications. Referencing the Ministry of Trade, Industry and Energy's AI Factory initiative, the report said the program "is planned around AI technology, but a company-centered perspective must be taken into account."

KIAT's core argument is that technology should be developed to meet the needs of individual businesses, rather than simply handing companies AI tools that have already been built. "The initiative is structured around integrating already-developed AI solutions into production processes," the report said, adding that "what matters on the ground is not finding which AI model or dataset fits a given process, but testing multiple applicable AI technologies against processes that need transformation and identifying solutions from there."

The report went on to say that "to uncover diverse AI application needs, a company-centered strategy is required — one that includes manufacturing AI consulting, development of process case studies, and a quantitative expansion strategy based on demand assessment."

The AI Factory initiative is a flagship manufacturing AI transformation project in which the ministry is investing 52.7 billion won ($39.4 million). Its aim is for companies to jointly collect process-specific data from operations where they want to introduce automation technology, and then use that data to develop AI models. The ministry announced the project last year and in April revealed 32 supported processes and industries.

KIAT also noted that the initiative's participation threshold is too high for small and medium-sized enterprises. "The project is structured around data sharing and joint AI model development among large, medium and small companies, but the subcontracting realities facing smaller firms and the cost burden that prevents them from securing AI talent limit their ability to participate actively," the report said.

KIAT recommended that support for manufacturing AI transformation be distributed evenly across businesses of all sizes, not just large corporations. Overseas, collaboration between large and small companies during AI transformation is common. The report cited German engineering firm Siemens as an example: the company offers free Internet of Things service solution accounts to other businesses, jointly collects data with them, and then applies the data it secures to its own software-as-a-service-based process automation.

To enable such a collaborative model in South Korea, the report called for policies tailored to smaller firms that struggle even at the data-collection stage. "Small and medium-sized enterprises lack the data infrastructure, collection centers and IoT devices needed to use manufacturing AI, and do not have sufficient engineers to manage them," it said, arguing that "efforts to improve productivity are needed to overcome the data collection infrastructure gap."

The report added that "a data-based value chain must be built by establishing a cooperative framework covering ownership of manufacturing data and profit-sharing under a data economy."