Artificial intelligence (AI) has emerged as a core keyword in the government's 2027 budget proposal. The significant expansion of the AI budget from 9.9 trillion won in 2026 to 21.3 trillion won in 2027 signifies that AI has become the central axis of national growth strategy and fiscal management. Now, the crucial question is not merely the scale of the budget, but what kind of qualitative AI transition this funding will create.
AI has established itself as an everyday tool in industrial settings, used for manufacturing defect detection, customer data analysis, and quality improvement. The government has also supported corporate AI utilization through vouchers, pilot projects, and R&D programs. In particular, the AI voucher program lowered entry barriers for small and mid-sized companies, provided market entry opportunities for suppliers, and contributed to initial diffusion.
However, it is time to change the question. We must move beyond asking "How much AI has been adopted?" to asking "How much have work methods and industrial structures changed due to AI?" Currently, two concepts are mixed within AX (AI Transformation) as used in policies and on the ground: 'AI Adoption,' which applies AI to existing work structures to improve efficiency, and 'AI-based Transformation,' which redesigns business processes and operational structures themselves based on AI.
These two concepts are not mutually exclusive but represent stages of development. Just as electricity evolved from lighting to transforming entire factory production systems, AI will evolve from an auxiliary tool into infrastructure that leads decision-making and execution. The problem is that the majority of current AX projects remain at the 'adoption' stage. AX in various fields often stops at adding AI modules to existing structures or merely attaching analytical functions.
There are reasons why projects have been designed with a focus on adoption. This reflects the characteristics of government finance, which must involve many companies and produce results in the short term, as well as the reality on the ground, where risks associated with changes in personnel, organization, and systems must be considered. However, now that we have entered the era of AI agents that independently find information, make judgments, and execute actions, it is difficult to drive fundamental industrial innovation through partial adoption alone.
In this regard, the Physical AI R&D project offers important implications. The Physical AI projects in Jeonbuk and Gyeongnam provinces, along with dark factory concepts, being promoted by the Ministry of Science and ICT, connect manufacturing sites, robots, data, AI models, and operational systems into one, going beyond simple automation. This is a representative case of an 'AI-native' approach that redesigns processes, equipment, and decision-making structures centered on AI, from an AI full-stack perspective.
AX policy must evolve in three directions. First, the government budget structure should be reorganized from a focus on short-term performance to one centered on industrial transition outcomes. Second, environmental preparation such as workforce retraining, organizational change support, and regulatory improvement must proceed in parallel. Third, there must be a shift away from individual task-centric approaches toward integrated programs that span data construction, technology development, piloting, diffusion, and operational internalization.
If AI adoption is the starting point, AI-based transformation is the direction we should head toward. The 21.3 trillion won AI budget should be used not merely to expand AI utilization, but to restructure the industrial landscape with AI as a premise. This is the core task for South Korea to leap into the top three in AI.
in OPINION
