Nam Hyun-woo has worked as a staff writer at The Korea Times since 2013, mostly covering business and politics. He currently belongs to the Business Desk where he covers topics such as emerging tech, AI, ICT and Korea's chaebol community. Prior to joining the team, he was the paper's correspondent for the presidential office of Korea during the Yoon Suk Yeol and Moon Jae-in administrations.
Utopia or dystopia? Governance, social consensus to shape future led by physical AI
Professor Lim Jong-in of Korea University's School of Cybersecurity, left, speaks with Superb AI CEO Kim Hyun-soo during the AMCHAM-Korea Times AI Forum 2025 at the Conrad Seoul, Tuesday. Korea Times photo by Shim Hyun-chul
By Nam Hyun-woo
Professor Lim Jong-in of of Korea University School of Cybersecurity speaks during the AMCHAM-Korea Times AI Forum 2025 at the Conrad Seoul, Tuesday. Korea Times photo by Shim Hyun-chul
As artificial intelligence (AI) expands from cyberspace into physical reality, the world is bracing for another profound transformation in the form of “physical AI,” a term that refers to AI systems that have physical forms that can interact with the real world.
During a fireside chat at the AI Forum 2025, co-hosted by The Korea Times and the American Chamber of Commerce in Korea (AMCHAM Korea), Tuesday, experts said they are certain that physical AI will transform industry, manufacturing, health care and every other sector of human life. At the same time, they stressed that whether the future turns into “a utopia or a dystopia” will depend not on the technology itself but on governance and social consensus on how to embrace the change.
“According to recent research, the digital AI market is projected to reach $1.8 trillion by 2030, but that of physical AI will amount to $50 trillion during the same period,” said professor Lim Jong-in of Korea University’s School of Cybersecurity.
“The next industrial revolution hinges on the completion of physical AI.”
Lim said the global race for physical AI has accelerated since Nvidia CEO Jensen Huang unveiled the company’s generative world foundation model, Cosmos, and AI training simulation platform, Omniverse, at CES 2025 in January. He added that the expanded use of synthetic data in AI training is further accelerating the development of physical AI.
Less than a decade ago, training AI relied mostly on human-created data. With advances in generative AI and synthetic data, however, AI has begun to train on artificial information created by algorithms, and the progress in simulation technology has enabled it to learn even from situations where real-world data is scarce.
Superb AI CEO Kim Hyun-soo speaks during the AMCHAM-Korea Times AI Forum 2025 at the Conrad Seoul, Tuesday. Korea Times photo by Shim Hyun-chul
“We have several clients in the shipbuilding and heavy industries sector, and the chronic challenge for physical video AI is those related to worker safety management,” said Kim Hyun-soo, CEO of Superb AI, which provides data labeling solutions and a vision foundation model called Zero.
“Detecting safety issues such as a worker collapse or fire requires training data, but such incidents rarely occur in reality. As a result, the lack of data on such events has hampered the performance of AI models. So we are now developing technologies that leverage synthetic data, creating virtual scenarios and data of such situations using generative AI.”
Kim noted that he used to be a skeptic of synthetic data due to the so-called simulation-to-real gap, but stressed that he is now “an ardent advocate” as the gap narrowed when AI models trained on simulated data are applied to actual worksites.
Lim noted that the progress of physical AI in the manufacturing sector could serve as a model for Korea-U.S. cooperation, forecasting that combining U.S. fundamental technologies with Korea’s manufacturing capabilities would create “unmatched competitiveness.”
At the same time, the two speakers cautioned that the introduction of physical AI could still spark public fears, as well as a decline in jobs for humans.
“Even video AI causes backlash such as privacy concerns at worksites,” Kim said. “And resistance is expected to be even greater if AI-powered robots are deployed.”
Both Lim and Kim stressed that social acceptance and AI literacy education must go hand in hand with the progress of physical AI, and a proper governance system is required to capitalize on any opportunities the technology will bring.
“Physical AI will ultimately lead to the development of humanoids and even surgical AI for autonomous vehicles,” Lim said. “In such cases, errors or hacking could cause serious human and corporate losses. That is why we need human-in-the-loop systems, stop mechanisms and international standards.”
Kim also said the governance for physical AI “should be stronger” than existing security measures, since robots can cause direct harm. “There must be standards to ensure traceability and clarify accountability in the event of a malfunction,” he said.