“Dokdo is a disputed territory between South Korea and Japan.”
Early generative artificial intelligence (AI) services initially responded this way when asked about Dokdo’s ownership. Although such answers disappeared as AI systems evolved and accumulated information, the initial shock was considerable.
The AI should conclude that Dokdo is a South Korean territory based on objective historical, geographical, and international legal evidence. The initial incorrect answers stemmed from biased training data. This issue is a key concern for latecomers as the AI competition between the U.S. and China intensifies.
This is also why it is necessary to develop independent AI ecosystems that reflect diverse cultures and histories. At a time when the AI ecosystem needs to expand in a positive direction, like civilization and life evolution, if we fail to contribute to diversity, we will inevitably fall behind.
Stanford University reported on Tuesday that private AI investment in the U.S. reached approximately $67.2 billion in 2023, followed by China at about $7.7 billion.
In contrast, South Korea’s AI investment was around $1.3 billion, trailing behind the UK ($3.7 billion), Germany ($1.9 billion), Sweden ($1.8 billion), France ($1.6 billion), Canada ($1.6 billion), and Israel ($1.5 billion).
While investment size isn’t the sole indicator of competitiveness, other metrics also signal a crisis. The Boston Consulting Group (BCG) didn’t classify South Korea as an AI leader; instead, it categorizes it as an “AI stable competitor,” indicating that Korea is somewhat lagging.
This category includes Japan, Malaysia, Spain, Taiwan, and Australia. South Korea, long hailed as an IT powerhouse, is now grouped with Japan, a country traditionally seen as weaker in digital technologies in the AI field.
This reality starkly contrasts with the South Korean government’s claim that it is one of the top three AI leaders. The country lacks investment and infrastructure, and there is also no clear AI strategy. This year, enacting a fundamental AI law containing support and regulatory measures for the AI industry is virtually impossible.
South Korea can’t keep up with the leading nations with massive investment bases. While companies like Naver, Kakao, SKT, KT, and LG U+ are developing Korean-language specialized AI, private sector efforts alone are insufficient to secure a competitive advantage.
Misunderstandings about AI sovereignty, viewing it in an exclusive sense, are also impeding innovation. Achieving digital sovereignty is a byproduct of developing various independent models that make up the AI ecosystem.
AI diversity isn’t about protecting our assets but leveraging our strengths to prepare for the future. When each country develops its own independent AI diversity, global ecosystem stability improves.
An IT industry insider warned, “If Anglo-American AI companies dominate the entire AI ecosystem, the world will be left with AI systems sharing a single set of values. Just as reduced genetic diversity is the main cause of biological extinction, the AI ecosystem should evolve by intertwining diverse, independent AI systems.”
To achieve this, countries must independently develop areas with a comparative advantage based on their culture, language, and history while enhancing capabilities in weaker areas through global collaboration.
The insider also noted, “The UK and Singapore are rapidly developing independent AI models, while Germany and France are expanding their influence across Europe. Government intervention is crucial since private companies alone can’t sufficiently support AI diversity. This could include bulk purchases of key resources like GPUs, tax incentives, and direct investment support.”
