Naver CEO Choi Suyeon: "Sovereign AI Isn't Exclusive Self-Sufficiency... Data Decision-Making Rights Are Key"

19th APK 'Industrial AI Tech Talk
"AI models are consumables... The gap comes from on-the-ground data"

"The essence of sovereign AI lies not in exclusive self-sufficiency, but in sovereignty over and control of data."

Choi Suyeon, CEO of Naver, gave this answer on the 31st during a discussion at the 19th Asia-Pacific German Business Conference (APK 2026), held at Grand InterContinental Seoul Parnas, when asked whether sovereign AI meant independently building every part of the value chain.

Choi Suyeon, CEO of Naver (center), speaks during a discussion at the 19th Asia-Pacific German Business Conference (APK 2026), held at the Grand InterContinental Seoul Parnas on the 31st. From left: moderator Jennifer Moon, Choi, and Roland Busch, chairman of the Asia-Pacific Committee of German Business (APA) and CEO of Siemens. Photo by Kim Jinyoung

Sovereign AI refers to the ability to develop, control, and operate AI directly, without being dependent on big tech companies, using a country's own infrastructure, data, and technical workforce. South Korea is pursuing "full-stack self-reliance," spanning infrastructure to models, based on generative AI models developed by private companies such as Naver and LG and an AI semiconductor ecosystem.

"It is nearly impossible for a country or company to develop 100% of everything itself, from AI semiconductor chips to foundation models and end-user applications," Choi said. "It is a natural progression for each party to build on its strengths and work with global partners." However, she emphasized, "Even amid this kind of cooperation, it is important to retain the right to make decisions about data and the ability to control it."

As AI evolves beyond simply providing answers to become an agent that performs actual work, Choi also defined the "boundaries of human intervention." "The results AI produces can look so impressive and logical that it is easy to assume that 'the problem has been perfectly solved,' but that is often not the case," she said. "No matter how logical the explanation may be, humans must separately check whether the problem we originally set out to solve has actually been resolved."

Choi continued, "Before introducing AI into physical settings, we must clearly define the scope of its use and put safeguards in place so that humans can immediately stop it and restore the previous state if it goes beyond the scope of control." She added, "Responsibility, too, should be allocated in line with control authority."

Roland Busch, chairman of the Asia-Pacific Committee of German Business (APA) and chairman of Siemens, also saw companies' own data, rather than AI models themselves, as a source of competitive advantage.

"From a company's perspective, an AI model itself is much like a consumable that can be swapped out for another at any time," Busch said. "Ultimately, the quality of data and domain expertise create the gap between companies." He added, "Just as Siemens is working with HD Hyundai to fully represent ships and shipyards in a 'digital twin,' AI must be controlled within the boundaries of strict physical laws."

Asked what headline she expected to see on the front page of a morning newspaper four years from now, Choi replied, "I hope the word 'AI' will no longer appear anywhere in the headlines." She added, "I hope AI will quietly permeate our daily lives and industries as a reliable part of the background, like air and electricity."

Busch said, "I dream of seeing these headlines on the front page: 'Industrial AI becomes the world's new operating system, taking global economic productivity to the next level' and 'Eradicating incurable diseases and ending hunger around the world.'"

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