Land minister meets experts to map out Korea's autonomous-driving AI ecosystem

Minister of Land, Infrastructure and Transport Kim Yun-deok met with autonomous-driving AI experts Tuesday to discuss ways to build a Korean-style innovation ecosystem for autonomous-driving AI.

The meeting came as the ministry works to develop Gwangju into an autonomous-driving demonstration city to generate the large-scale, high-quality training data needed for autonomous-driving AI development, and to pursue AI training data standardization for end-to-end technology development in cooperation with the Ministry of Science and ICT. Kim also heard firsthand accounts of challenges encountered during autonomous-driving AI research and development.

"Autonomous driving is a flagship physical AI industry that combines automobiles and AI, and the key to the autonomous-driving industry lies in how much diverse training data can be secured and how effectively AI models can learn from it," Kim said. She added that the autonomous-driving demonstration city is "a project to secure large-scale, high-quality data that the private sector finds difficult to build on its own, and to provide AI training infrastructure and a demonstration environment." She stressed that building a national-level data flywheel ecosystem — one that shares secured data to feed AI model development and then uses demonstration to improve performance — is of paramount importance.

Participants at the meeting, including Seoul National University professor Choi Jun-won, Kakao Mobility Executive Vice President Kim Jin-gyu and Ride Flux Co-representative Jeong Ha-wook, said the competitiveness of autonomous-driving AI is determined by the scale and diversity of data secured from real road environments through data scaling. They proposed that to develop autonomous-driving AI capable of operating safely anywhere in the country, data must be continuously gathered from diverse regions and road environments starting with the Gwangju demonstration city, and an open ecosystem must be built that allows industry, academia and research institutions to use it.

In response, Kim said the ministry would build an autonomous-driving AI cluster covering the full cycle from AI model research and development through demonstration to commercialization, securing high-quality training data across more cities, and in doing so create an innovation ecosystem to lead the advancement of physical AI technology in line with the nation's three mega projects for a great leap forward.