National AI Strategy Committee Holds Speed Adjustment Seminar: "Accountability Needed Over Slowing Down"

"Speed with Accountability Required
for Advancing Both Frontier AI and Security Capabilities"

Amid rising calls for slowing the pace of artificial intelligence (AI) development in the global AI industry, a forum was held to discuss Korea’s response. As AI continues to advance rapidly, raising concerns that it may develop beyond human control, experts have proposed a “speed with accountability” approach. This strategy entails not only fostering Korea’s AI to a frontier-level but also strengthening security capabilities in parallel.

Lee Wontae, Chair of the Security Subcommittee at the National AI Strategy Committee, emphasized at a seminar on AI development speed adjustment held on October 1, "Korea must simultaneously and rapidly secure frontier capabilities as well as safety, security, and control competencies, in line with the principle of 'speed with accountability.'" Delivering the keynote speech, he presented an overview of the major countries’ approaches to moderating AI development speed.

Ha Jung-woo, Vice Chairman of the National Artificial Intelligence Strategy Committee, is speaking at the 13th Science and Technology Ministers' Meeting held on the 30th at the Main Video Conference Room of the Government Seoul Office in Jongno-gu, Seoul. Photo by Yonhap News Agency

Choi Daeseon, Director of the AI Safety Research Center and professor at Soongsil University, pointed out in his presentation that the current approaches—focusing on high-impact AI in practical applications or computationally intensive cutting-edge AI—are inadequate for addressing the risks of frontier AI, which is expected to expand rapidly in scope and unpredictability. He insisted that both technical and regulatory countermeasures need to be reviewed.

Within the industry, priority was placed on achieving frontier-level AI performance. Sung-hoon Kim, CEO of Upstage, said, "AI is moving beyond superintelligence (SI) into an era of 'self-evolving AI.' However, adjusting the domestic AI development speed prematurely may lock in the technology gap with global big tech and increase dependence. Therefore, Korea should first secure self-evolving technology, and only then consider implementing safety mechanisms and speed adjustments in line with global trends."

Hanul Lee, Vice President at AIM Intelligence, proposed that Korea should simultaneously pursue independent model evaluation and control capacities, as well as international cooperation.

The session was followed by a panel discussion and on-site Q&A with the presenters, Myungjoo Kim, head of the AI Safety Research Institute, and Donghwan Kim, CEO of Forty Two Maru. The discussion covered strategies for implementing “speed with accountability” considering major countries’ policies and Korea’s domestic environment, as well as support measures for closing the technology gap with global big tech and strengthening domestic AI competitiveness.

Ha Jungwoo, Standing Vice Chairman of the National AI Strategy Committee, emphasized, "Although there are diverse concerns and public discourses regarding slowing down AI development, both domestically and internationally, securing safe frontier AI capabilities is not optional, but essential. Especially during this period of dramatic change, the government and private sector must come together resolutely to advance toward our goals."

This content was produced with the assistance of AI translation services.