The government and companies have unveiled a blueprint to invest more than 1,000 trillion won in building 18.4GW-class artificial intelligence (AI) data centers by 2035, and to commercialize humanoid robots across 10 major industries by 2028. Yet a strategy to secure the security personnel needed to operate them safely has failed to keep pace. When AI robots are deployed at factories, logistics sites, shipyards, and defense facilities, security incidents can go beyond simple data leaks to cause equipment damage and casualties. Even so, the system for cultivating convergence personnel who understand robot security, manufacturing operational technology (OT), and AI model security together remains in its early stages. Voices are growing that Korea must move quickly to secure robotics security personnel equipped with both information security capabilities and an understanding of robot systems.
According to the domestic security industry on Wednesday, concerns over a shortage of security personnel have grown recently in hiring. Robotics security requires expertise in information security, robotics, and AI all at once. Personnel must understand robot software frameworks such as robot operating systems (ROS, ROS2), industrial control systems (OT), and industrial security standards, as well as have the ability to analyze complex threats including AI model attacks, sensor disruption, and communication network intrusion. This is because when robot systems are attacked, the consequences can go beyond data leaks and service disruptions to production line stoppages, equipment destruction, and worker safety accidents.
However, the industry's assessment is that finding convergence talent equipped with such capabilities is like "picking a star from the sky." An official at cybersecurity firm AhnLab said, "Personnel with robot security capabilities are currently being secured mainly around some robot companies and security organizations, but there are still not many specialized organizations dedicated to robot attack research or vulnerability analysis." He added, "In particular, personnel who understand robot software, industrial security, embedded systems, and AI technology together are in short supply across the industry, so competition to secure talent appears likely to intensify further."
An official at another security company also said, "In the wake of a series of hacking incidents last year, demand for security-enhanced software developers has already grown unusually, making it difficult to find experts." He added, "Compared to demand for physical AI, related security personnel are scarce worldwide, so we will have to compete with overseas companies for hiring going forward." In fact, global research firm Future Market Insights projected that the global robotics security market would grow roughly threefold, from $4.7 billion (about 6.3 trillion won) last year to $14.3 billion (about 19.3 trillion won) by 2035.
At the government level, measures specialized for robotics security have not yet been made concrete. The government has unveiled a comprehensive plan to cultivate AI robot personnel, but has failed to prepare a separate specialized track encompassing robot security, OT security, and AI security. The second Physical AI Alliance, launched last month, also has a talent cultivation action group under its "foundational governance" subcommittee, but a detailed roadmap for cultivating robotics security personnel has not yet been established. Lee Da-som, a KAIST professor who heads the talent cultivation action group, stressed, "Because robotics security itself is such a nascent field, industry, academia, and research institutes need to actively cooperate from now on to discuss cultivation plans." She added, "In the physical AI field, securing the personnel who can handle it is as important as technology or standard security." She continued, "In the short term, we should take an approach of connecting and retraining security personnel and robotics personnel, and in the mid-to-long term, we need a specialized track that raises convergence talent from the start."
In the global market, security requirements related to AI robots are also being strengthened. A representative example is the European Union's (EU) Cyber Resilience Act (CRA). The CRA is a regulation requiring cybersecurity to be applied from the design and development stages for "products with digital elements (PDE)," with vulnerability and incident reporting obligations taking effect this September. The United States is also pushing the "Cyber Trust Mark" program, which grants a security certification label to consumer Internet of Things (IoT) products that meet the security standards of the National Institute of Standards and Technology (NIST). Although their scope and enforceability differ, both systems reflect how the security reliability of digital products is emerging as a certification and purchasing criterion in the global market.
Accordingly, experts stress that to preempt the global robot export market, the government must quickly establish concrete guidelines related to robotics security, and companies must begin security design as soon as possible. Lim Myung-chul, head of the IoT Convergence Security Research Institute at cybersecurity specialist Penta Security, warned, "If the government focuses only on boosting the performance of domestic AI robots and misses security design, there is a high risk of setbacks in the future process of expanding overseas." He advised, "From the robot design stage, we need to internalize security from a 'lifecycle' perspective, going beyond simple hacking prevention."
