뉴스

U.S. Official: 'Complete Domestic Ownership of AI Hinders Progress; Must Not Discriminate Against U.S. Firms'

U.S. Official: 'Complete Domestic Ownership of AI Hinders Progress; Must Not Discriminate Against U.S. Firms'
안내

We only offer this video
to viewers located within Korea
(해당 영상은 해외에서 재생이 불가합니다)

▲ Russ Hadley, Senior Advisor at the U.S. Department of State's Bureau of Cyberspace and Digital Policy, speaks at a seminar hosted by the National Bureau of Asian Research (NBR) on "Digital Infrastructure Security in South Korea's AI Sector" in Washington D.C. on June 23 (local time).

A U.S. State Department official argued that the concept of "AI sovereignty," currently being discussed in South Korea and other nations around the world, should not entail full domestic ownership of the entire AI stack, but rather focus on securing appropriate control systems.
Russ Hadley, Senior Advisor at the State Department's Bureau of Cyberspace and Digital Policy, expressed this position at a seminar hosted by the National Bureau of Asian Research (NBR) in Washington D.C. on June 23 (local time), titled "Digital Infrastructure Security in South Korea's AI Sector."
Hadley noted that many countries tend to view AI sovereignty as "complete domestic ownership of the entire AI stack, including chips, data, models, and infrastructure." He stated that "the AI stack is inherently very complex, and attempts to build it domestically from scratch are not only incredibly costly but also risk leaving that country behind while technology advances rapidly elsewhere."
He cited "unnecessarily restrictive data localization requirements, broad restrictions on foreign cloud providers, network usage fees targeting content platforms, and procurement rules that prevent governments from selecting the most capable technology providers" as examples of excessive AI sovereignty.
He continued, "In South Korea, the rapid advancement of AI has been met with barriers such as requirements for physical separation of government servers and comprehensive data localization policies," arguing that "this poses significant risks to South Korea itself."
He further emphasized, "In reality, these measures increase costs, reduce competition, and in many cases, actually increase security risks," adding, "What the U.S. cannot accept in this area are appeals to digital sovereignty designed to exclude or discriminate against U.S. companies."
Hadley added, "Digital sovereignty should mean verifiable control, not physical possession," and suggested that "shifting to modernized regulations that allow for logical server separation (rather than physical) and the cross-border flow of low-to-medium-sensitivity data would enable South Korea to leverage the benefits of AI in the public sector, particularly in cyber defense."
He also noted, "These are priorities that were publicly listed in the U.S.-ROK joint fact sheet last November."
Hadley argued that U.S. AI technology firms "can build large-scale, independent infrastructure with secure supply chains," and that the best path for AI development is for "(South Korea's) leading companies to rapidly adopt the best technology stacks currently available (by partnering with U.S. firms) while developing their own technologies."
"In reality, there is no one doing it better (than U.S. firms)," he said, adding that instead, "sensitive data remains within your borders, and your policy choices remain your own—that is how we define (AI) sovereignty."
(Photo: Yonhap News)
※ Please note: This article was translated by AI and may contain errors.
Copyright Ⓒ SBS & SBSi. All rights reserved.
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.

Most Read