"We will actively support domestic companies in verifying and commercializing specialized networks and next-generation AI network technologies by linking demonstration environments with R&D," said Lee Sang-hoon, head of the Korea Broadcasting and Telecommunications Commission (KCA).
AI is no longer limited to creating broadcasting content. It now manages communication networks and operates factories and logistics. KCA, which was once a broadcasting and radio support agency, is expanding its role as an infrastructure institution for industrial AI.
What does AI need to function effectively? Most people think of GPUs and data centers. However, even the smartest AI is hindered if the pathways for data are blocked. Just as the human brain uses nerves to move the body, AI requires a network to receive information and transmit commands. Now, AI is beginning to inhabit that neural network.
A prime example is the 'Next-Generation Network AI Foundation Model Development' project, led by the Electronics and Telecommunications Research Institute (ETRI) with KCA, KT, and HFR. This national R&D initiative aims to create a Network AI Foundation Model (NFM) by 2030. While generative AI learns human language, NFM aims to understand and operate networks by learning from the data generated within communication networks.
KCA's role is hands-on. It involves collecting log data from 5G specialized networks to build a data lake, refining various multimodal data, and deploying AI agents in actual specialized network operations for validation. This process is not just about creating models in a lab; it involves running them on live networks.
The key to this transformation lies in operational methods. Traditionally, communication networks have relied on human oversight to monitor conditions and troubleshoot issues. If AI can read vast amounts of network data in real-time, it can detect anomalies, identify causes of failures, and assist in operations. We are transitioning from an era where AI uses networks to one where AI manages networks.
This shift is particularly relevant for 5G specialized networks, which differ from public mobile networks. These dedicated networks are tailored for factories, logistics centers, hospitals, and agricultural facilities. In environments where numerous machines, sensors, and robots connect simultaneously, fast and reliable communication is not optional; it is essential.
The potential of this technology is exemplified by the Jeonnam Muan Southwest Vegetable Cooperative's Agricultural Production Center (APC). KCA has installed a 5G specialized network there, automating the process of sorting, packaging, and shipping agricultural products using 5G, robotics, sensing, and autonomous driving technologies.
In the AI era, communication networks must be open. This is where open RAN becomes crucial. Open RAN technology transforms networks into an open structure, allowing equipment and software from different manufacturers to work together. This breaks the dependency on specific corporate equipment and opens the door for more companies to enter the network ecosystem.
KCA is assisting companies in testing and validating specialized network devices and services at the 'I-eum 5G Testbed' in Iksan, North Jeolla Province. This platform allows companies with new technologies to test their products without having to build a network from scratch.
The role of public institutions in the AI era is critical. Developing technology alone does not automatically create an industry. It is essential to verify whether successful technologies from the lab can operate in factories and logistics centers, enabling companies to turn them into marketable products.
There exists a 'valley of validation' between technology and market. Small and medium-sized enterprises and startups with good technology often struggle to cross this valley due to the high costs of creating test environments that resemble industrial settings. This is why KCA must connect testing grounds and R&D to support the commercialization of businesses.
As networks become smarter, the importance of security increases. In a scenario where factories, robots, and logistics centers are interconnected, a breach in the communication network can lead to more than just slow internet; it can halt production lines and disrupt robots and autonomous equipment.
KCA's involvement in developing identification and testing technologies for 5G specialized network devices based on quantum security is crucial. As we enter the era of quantum computing, existing encryption systems will face new challenges. It is essential to integrate new security technologies, such as post-quantum cryptography (PQC), into networks and validate them in real-world environments.
AI, 5G, open RAN, and quantum security may seem like distinct technologies, but they operate as a cohesive unit in industrial settings. For AI to make decisions, data must flow, and for data to flow, a network is necessary. This network must incorporate equipment from various companies, and the entire system must be secure. If AI is the brain, the network is the nervous system, and security is the immune system.
Broadcasting is no exception. The boundaries between broadcasting, telecommunications, and radio, once distinct, are rapidly dissolving in the AI era. KCA is supporting AI and digital-based broadcasting program production, encouraging the use of AI throughout planning, production, and distribution.
Generative AI significantly streamlines many aspects of the production process, including research, video and image creation, subtitles, and translation. This presents a unique opportunity for small broadcasters and production companies. Content that once required substantial budgets and personnel can now be created by smaller teams. If translation and dubbing costs are reduced, domestic small production companies can target global markets from the outset.
However, KCA's AI transformation should not be confined to broadcasting support. Network AI, 5G specialized networks, open RAN, quantum security, and AI in broadcasting content all fit into a larger picture: laying the groundwork for an AI-driven industry. The national infrastructure in the AI era will not be limited to data centers; it will connect numerous spaces, including manufacturing plants, logistics centers, agricultural facilities, and broadcasting production sites through AI networks.
This is why Lee Sang-hoon emphasizes the connection between demonstration environments and R&D. It is not enough to create a technology with government research funds; companies must test that technology in real environments and pave the way for entering international markets. The future of KCA, which has managed radio waves and supported broadcasting, lies here.
KCA aims to become a platform that connects technology companies, broadcasters, telecommunications firms, manufacturers, research institutions, and markets, facilitating the integration of AI and network technologies into industrial settings.
The competition in AI will not be determined solely by who creates the best model. The country that successfully implements that AI in factories, logistics, broadcasting, and urban environments will emerge victorious. The effort to establish an 'AI neural network' in South Korea's industry is where KCA's AI transformation, led by Lee Sang-hoon, begins.
Lee Sang-hoon, head of KCA, is an expert in ICT administration who has dealt with policies in the fields of radio, broadcasting, information technology, and information security. He served as the Director of Multilateral Cooperation at the Ministry of Science and ICT, Director of Information Security Planning at the Ministry of Science and ICT, Director of the Postal Business Information Center, and Head of the Central Radio Management Office before being appointed as the 8th head of KCA in November 2023. He is leading the promotion of the broadcasting media industry, the establishment of a radio utilization environment, and the validation and industrial expansion of next-generation ICT technologies such as 5G specialized networks, network AI, and open RAN.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.