"Who stops the Terminator in the movie? It is Arnold Schwarzenegger (the older model Terminator). Without him, humanity would be immediately annihilated."
Kim Seong-hoon, CEO of Upstage, argued at a seminar on "AI Development Speed Regulation Theory" held on the 1st in a conference room on the 16th floor of Seoul Square, hosted by the National AI Strategy Committee, that "to overcome SI (Superintelligence) that is not on our side, we must have our own SI," asserting that AI development should continue.
The seminar was organized to consider Korea's response direction amid the rising debate on regulating the pace of AI development centered around U.S. big tech companies. Many industry, academic, and research attendees present also voiced a unified call for developing Korea's AI technology safely and rapidly to strengthen AI security.
Ha Jeong-woo, Executive Vice Chairman of the AI Strategy Committee, emphasized in a video congratulatory message that "while it is true that AI risks are increasing, it is important to strengthen evaluation systems and capabilities regarding what mechanisms to create and how to develop safely," adding, "To take a leading voice in safety regulations, we must build AI well, have influence, and possess power." This means that technological capability must be secured first to gain influence in global discussions on AI safety norms.
Lee Won-taek, Chairman of the Security Special Committee of the AI Strategy Committee, evaluated the "Speed-Responsibility Theory" proposed by Deputy Prime Minister and Minister of Science and ICT, Baek Gyun, as a direction Korea should take. Lee (Chairman) stated, "What the CEO of Anthropic said was not to stop AI development, but to secure time for control," adding, "Korea's answer is 'speed-responsibility': do not delay frontier models or industrial applications, but increase accident recovery capability and responsibility speed at a pace faster than the rate at which risks grow."
He further stated, "Cutting off AI authority in the event of an accident to ensure controllability is the path to finding true AI sovereignty," adding, "In the gap between the U.S.-China bipolar structure, it is necessary to engage in functional solidarity and formulate an AI diplomacy strategy as multiple countries pursue multilateral cooperation."
Warnings were also issued about preparing for imminent risks caused by AI. Choi Dae-sun, Director of the AI Safety Research Center at Soongsil University, noted that simply combining agents with publicly available AI models can yield cyberattack capabilities surpassing frontier models, stating, "We are not a frontier player, so we cannot say it is not our concern."
Concerns emerged from the industry that slowing down could deepen technological dependence. Kim Seong-hun (CEO) referenced the agreement at the U.S.-China summit for both countries to use the term SI (Super Intelligence) instead of AI, stating, "The call to stop may be a message from places capable of creating SI, like the United States and China, running at full speed, directed at countries below them," adding, "If we also stop together, we will crawl on the ground; we too must reach SI."
Opinions were also raised that sharing various accident cases and experiences is important to prevent security risks caused by AI. Lee Han-eul, Vice President of AIM Intelligence, stated, "First, creating top-tier models is essential, and sharing safe utilization experience for AI security research and incident information is crucial," adding, "Since we all have vulnerabilities, we should boast about and share how we overcame them."
in TECH
