The conversation between President Lee Jae-myung and Demis Hassabis at Cheong Wa Dae on April 27 was brief but pointed. When Lee asked whether universal basic income would be needed in the AI era, Hassabis agreed, saying society would have to redesign how it defines work and redistributes wealth. He also said artificial general intelligence, or AGI — systems that can match all human cognitive abilities — could come into view within the next five years. The remark was framed not as distant speculation but as a diagnosis of change already underway.
Hassabis is not known for exaggeration. He led breakthroughs such as AlphaGo and AlphaFold, and his warning focused on pace: AI is advancing faster than institutions can be updated and societies can adapt. Competition has begun and will not stop, leading many to conclude AI itself cannot be controlled.
The article argues, however, that stopping at that conclusion misses a key point. While the speed of AI may be hard to slow, governments can still shape the environment in which it operates — including rules for data flows and use, and how markets function. The central task, it says, is not simply to chase technology but to set the rules that govern it, the one area where individual countries can retain initiative.
It says the global AI race is shifting from company-to-company competition to a contest over national order. Big Tech firms such as Google, Microsoft and OpenAI are consolidating ecosystems with capital and data, making it unrealistic for latecomers to compete on identical terms. The proposed alternative is to design rules and build a structure in which competition takes place within them.
From that perspective, the article says Hassabis’ visit to South Korea and plans to establish an AI campus should be viewed with both opportunity and risk. Attracting overseas research hubs can support joint research and technology exchange, but it warns that global Big Tech does not readily share core technology and can instead absorb talent and data as it expands. A campus, it says, can be a site of cooperation and a channel for talent outflow.
It calls for conditional cooperation: ensuring some joint research results accrue domestically, creating safeguards against one-way talent leakage, and keeping data use within domestic norms. The piece describes this as a new form of industrial diplomacy — open to collaboration while protecting initiative.
The AGI debate, it adds, should be seen in the same context. If technology emerges that can replace human cognitive abilities, labor markets will be shaken, but that does not mean human roles disappear immediately. It outlines a transition: in the short term, humans continue to design and deploy AI; in the medium term, human-machine collaboration becomes common; in the long term, replacement occurs in some areas.
The priority now, it argues, is not fear but building the capacity to work with AI. Skills such as data interpretation, algorithm design and system integration remain important — not as an end state, but as capabilities needed to navigate the transition. Talent strategy, it says, should distinguish between short- and long-term needs.
The article says the larger issue is distribution. AI can sharply raise productivity while accelerating wealth concentration, with firms that control data and algorithms capturing most gains. It says Hassabis raised basic income in that context, but argues basic income is less an answer than a question because the central issue is how to fund it.
It points to limits of national policy as Big Tech earns profits across borders, making it difficult for any one country to capture and tax those gains. It proposes two approaches: international coordination, including global taxation systems such as a digital tax to reclaim excess profits; and domestic systems that return value for data use and ensure AI-generated gains circulate back to society.
Companies must also change, it says. Manufacturers such as Samsung Electronics and SK hynix should treat AI not as a tool but as core competitiveness, and the same applies to autos, shipbuilding and energy. AI, it argues, is a foundation for all industries, and failure to use it effectively could quickly erode existing strengths.
In the end, it says competition in the AI era comes down to three capabilities: developing technology, using it, and making the rules under which it operates. South Korea has some strength in the first and is catching up in the second, but still has far to go on the third.
The piece concludes that Hassabis’ remarks were not merely a technology forecast but a call to choose. If AI’s speed cannot be stopped, it argues, countries must design the world in which AI operates — and decide whether to be swept along or to build order on top of rapid change. It says South Korea now faces a crossroads between remaining a follower and becoming a rule-maker, with the remaining question being execution.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.