"To effectively promote M.AX, it is crucial to realize robot transformation through 'Accelerating AX, Realizing RX.'"
This is the direction proposed by Cho Young-hoon, head of the Korea Robotics Industry Promotion Agency. While generative AI has ushered in an era of AI that speaks and thinks, physical AI refers to AI that moves and works directly in the real world. Cho is focusing policy efforts on linking the AI transformation in manufacturing to robot transformation, bringing lab technologies into factories, logistics, and defense sectors.
AI is beginning to emerge in the real world.
Generative AI, symbolized by ChatGPT, has opened a new era of writing, drawing, data analysis, and human interaction. However, most of the AI we have experienced so far has operated within computers and smartphones. The next competitive stage is shifting to the real world, where AI recognizes its surroundings through cameras and sensors, makes autonomous decisions, and operates robotic arms and legs to assemble parts in factories, transport goods in warehouses, and perform dangerous tasks that are difficult for humans.
This is what is known as physical AI. This change holds special significance for South Korea, which has a global manufacturing base in automobiles, semiconductors, batteries, shipbuilding, and electronics. By combining AI as the brain with robots as the body, South Korea can create new industrial competitiveness distinct from generative AI.
Cho's emphasis on the transition from 'AX to RX' is precisely at this point. He aims to connect the AI transformation in manufacturing to the actual transformation of robots in industrial settings.
"Accelerating AX, Realizing RX"
The phrase 'Accelerating AX, Realizing RX' succinctly captures Cho's policy direction. It signifies the need to accelerate AI transformation and implement it as actual robot transformation. It is essential not only to develop AI technologies but also to ensure they operate effectively in industrial and public settings.
The competitiveness of physical AI is not determined solely by performance in papers or labs. It must be verified how accurately robots perform tasks in actual factories, whether they can work alongside humans in logistics, and if they can reliably carry out missions in hazardous environments.
The role of the Korea Robotics Industry Promotion Agency is also crucial here. It aims to bridge the gap between technologies developed in labs and the market through demonstration, certification, commercialization, and connecting with demand sources. Enhancing the completeness of technology is as important as making it usable in the market.
The military becomes a proving ground for physical AI.
A notable area in Cho's physical AI strategy is defense. On August 7, the Korea Robotics Industry Promotion Agency held a meeting at its Daegu headquarters with the Army's 2nd Operations Command to discuss expanding the application of civilian robot technologies in military logistics and supply operations.
In a context of declining troop resources due to population decrease and increasing burdens on military logistics, the agency believes that assigning repetitive tasks to robots can enhance efficiency and reduce the risk exposure of soldiers.
At the meeting, Cho stated, "The 2nd Operations Command is proactive in adopting physical AI and robots to realize a technologically advanced military, and we will support the discovery of robot utilization needs within the military to ensure practical application on the ground."
The key is 'robots that are practically utilized on-site.' It is essential to confirm results by integrating robots into logistics, supply, repetitive tasks, and hazardous operations, rather than merely demonstrating technology. Defense can become a new market for civilian robot companies and a means to supplement the military's workforce shortages.
In the era of humanoids, safety is competitiveness.
The technology symbolizing the physical AI era is humanoids. A new era is approaching where robots resembling humans work using spaces and tools designed for people. However, industry cannot be built solely on flashy demonstrations. Safety and reliability must be verified before deployment in the field.
This is why the establishment of the 'Humanoid Robot Safety Certification Center,' overseen by the Korea Robotics Industry Promotion Agency, is significant. The city of Daegu plans to invest a total of 18.65 billion won over five years from 2026 to 2030 to build the center at the National Robot Test Field site in Daegu. The funding will come from national, local, and private sources, amounting to 9.68 billion won, 7.8 billion won, and 1.17 billion won, respectively, with the Korea Robotics Industry Promotion Agency overseeing the project.
The center will be equipped to evaluate the dynamic stability of robots, the reliability of AI, and cybersecurity all at once, utilizing the facilities of the National Robot Test Field to support testing, demonstration, and certification seamlessly. This initiative aims to assist domestic companies in entering the market without the burden of overseas certification, in line with the European Union's trend of strengthening regulations through AI and cybersecurity laws.
The competition in physical AI is not just about who creates robots first. It is also about who can bring safer and more reliable robots to market.
Good technology alone is not enough for companies to survive.
Another challenge that Cho emphasizes is the commercialization of robot startups. While they have technology, they often lack places to validate their products, making it difficult to find customers, and without customers, securing investment becomes challenging. This issue is particularly pronounced in the robot industry, where development periods are long and investment costs are high.
On July 30, the Korea Robotics Industry Promotion Agency held a meeting with seven robot startups at its Daegu headquarters to discuss challenges related to funding, securing demonstration spaces, and developing initial sales channels. Attending companies expressed concerns about the lack of opportunities to validate product performance in the market despite developing excellent technologies.
Cho remarked, "Many startups are halting their ventures or leaving the robotics industry due to difficulties in the commercialization process, despite having good technology and experience. We will actively support startups in achieving tangible commercialization results through PoC, business matching, and investment attraction."
This signifies a shift in policy focus from technology development itself to actual commercialization outcomes.
A growth ladder from the lab to the market.
In the era of physical AI, the growth ladder between technology development and the market becomes even more critical. While software can be developed and tested at relatively low costs, robots require the production of hardware and the integration of sensors, motors, reducers, and control systems. Safety and durability must also be verified over extended periods in real-world environments.
Therefore, it is not easy for startups to manage the entire process from technology development to demonstration, certification, and market entry alone. This is why national infrastructure is needed to provide testing fields and demonstration spaces and to connect with demand companies to create initial markets.
The structure the agency aims to create does not involve jumping directly from technology development to the market. It involves a phased approach where concepts are validated in demand sources, performance is confirmed, and trust is built through demonstration and certification before connecting with demand companies and investors. This process must function effectively to help robotics companies overcome the 'valley of death' where they have technology but cannot enter the market.
Physical AI represents a new opportunity for South Korea's manufacturing industry.
While major tech companies in the U.S. and China lead the generative AI competition, the conditions for success in physical AI are somewhat different. Exceptional AI models alone are not sufficient. There is a need for robots and components to operate the AI, manufacturing capabilities to produce the robots, and industrial sites to deploy, learn, and validate the robots.
In this regard, South Korea is seen as having an opportunity. The country already has world-class manufacturing companies, numerous factories, and established industries in semiconductors, batteries, automobiles, and machinery. By combining AI as the brain, South Korea can expand its manufacturing competitiveness into the realm of physical AI.
Moreover, if the proving grounds are expanded to logistics, defense, and public services, it is possible to utilize the entire country as a single physical AI testbed. Technologies developed in labs must be repeatedly validated in various fields to gain recognition in the global market.
The results of AX must be confirmed in RX.
The meaning of 'Accelerating AX, Realizing RX,' as articulated by Cho, ultimately lies in the field. It is not just about the fact that AI has been introduced into manufacturing companies; what matters is how much AI has improved productivity and safety by operating robots on-site.
Humanoids that have demonstrated performance in labs must be able to work reliably in industrial settings, and startups with excellent robot technologies must meet customers after undergoing demonstration and certification. In areas like defense and logistics, where labor shortages and safety issues are significant, it is also essential to evaluate how much robots can reduce the burden on humans.
Thus, the tasks ahead are clear. We must create a tighter connection between AI research and development and manufacturing sites, and establish a structure that allows robots that have undergone demonstration and certification to quickly enter the actual market. For humanoids, speed alone will not suffice. It is crucial to verify safety and cybersecurity to produce products that the global market can trust.
While generative AI is changing human intellectual productivity, physical AI is beginning to transform the very workplaces where humans operate. If South Korea successfully transitions its manufacturing strength into the competitiveness of robots and AI, it can create new industrial opportunities.
The role of Cho Young-hoon and the Korea Robotics Industry Promotion Agency is also clear. They must become an institution that enables technology to operate in industrial settings, rather than merely showcasing it.
The path for South Korea to become a leader in physical AI will likely be determined not in labs, but in factories, logistics sites, defense, and service environments.
:Cho Young-hoon, Head of the Korea Robotics Industry Promotion Agency:
Cho emphasizes the connection between the AI transformation in manufacturing and the actual transformation of robots, focusing on expanding physical AI demonstrations in industrial and defense sectors, establishing a safety certification system for humanoids, and supporting startups in connecting with demand companies and securing investments to create an ecosystem where robot technology achieves results beyond the lab.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.