Lee Jong-won, head of the Humanoid Research Center at Korea Institute of Science and Technology (KIST), presents the project overview and direction./Courtesy of KIST

It has become difficult to stand out in the AI humanoid market by going it alone. Through this opportunity, we will achieve consolidation among researchers representing Korea and build a humanoid platform that represents Korea.

Lee Jong-won, Director General of the Korea Institute of Science and Technology (KIST) humanoid research division, explained the significance of the project at the kickoff meeting for the project to advance fundamental AI humanoid technologies based on public-private cooperation, held at KIST in Seoul on the 18th. The aim is to bind the capabilities of industry, academia, research, and medical sectors into one, given the judgment that it is difficult for individual institutions or corporations to keep pace with global competition by developing technologies on their own.

An AI humanoid is a technology that combines AI with a robot shaped like a human. It is a robot that perceives its surroundings, understands human speech, and plans and performs tasks on its own. Because it can be used in various spaces, including factories, logistics, hospitals, and care settings, it is regarded as a key field in the next-generation robot market.

Lee said, Overseas, with large-scale funding and personnel flowing into the AI robot industry, the competitive landscape is rapidly changing in terms of development speed and scale, adding, In contrast, domestically, the scale of bipedal humanoid production and specialized personnel is limited, making it difficult for individual institutions or corporations to compete alone.

To overcome these limits, the government designed this project as a national team approach that brings together domestic researchers and corporations. KIST is leading the project, with industry participation from LG Electronics, LG AI Research Institute, LG Energy Solution, Robostar, and WeRobotics. Academia, including Seoul National University, KAIST, Korea University, and Kyunghee University, as well as Hallym University Sacred Heart Hospital, are also joining. Through this, the government aims to secure a representative Korea-style AI humanoid platform by 2030.

Participants and officials of the public–private AI Humanoid Project kickoff workshop pose for a commemorative photo./Courtesy of Korea Institute of Science and Technology (KIST)

KIST President Oh Sang-rok said, China's rapid rise in competitiveness in advanced manufacturing was backed by public-private cooperation. This project was born from the idea that we should jump up as well, adding, We need to choose and focus on the technologies that require development and, with a one-team strategy, enter the humanoid market currently dominated by China and the United States.

The project will be run under three research divisions: physical AI, behavior software, and sensory hardware. The physical AI division will develop AI models that integrate visual, tactile, language, and behavior information so robots can understand and judge surrounding situations. In behavior software, the project will advance full-body control technology so humanoids can balance and move like people, handle objects, and continue long-duration tasks. On the hardware side, next-generation humanlike robot models will be developed based on KIST's humanoid platform KAPEX.

The developed technologies will be verified in real-world settings. The research team plans to develop more than 20 humanoids for joint use and aims to continuously perform at least two types of services for over one month in demonstration environments. In medical and care settings such as Hallym University Sacred Heart Hospital, they will test capabilities including daily living assistance, public services, and long-duration complex task performance. Because these robots move in spaces shared with people, the establishment of safety standards and certification systems will proceed in parallel.

Lee explained, If past R&D tended to stop at publishing papers or securing individual technologies, in this project, good papers are an interim goal, adding, The goal is to integrate the research achievements of each lab and institution into one and succeed in field demonstrations.

※ This article has been translated by AI. Share your feedback here.