Why South Korea Is Giving 51 Million People Unlimited AI

Before the end of this year, every resident of South Korea will have free and unmetered access to a general-purpose AI chatbot and a public service agent built on domestic models. The capacity to evaluate how those systems affect hiring, credit, and public services remains to be built, and it is the part of the plan where cooperation with the United States would matter most.

Korean President Lee Jae Myung set these plans in motion at the San Francisco AI Summit in July as part of his vision to transform Korea into an AI powerhouse, second only to the United States and China. His vision for achieving this goal rests on three pillars: Korea as an irreplaceable node in the AI supply chain through its dominance in memory chips; Korea as a testbed that adopts AI faster and more broadly than any other country; and Korea as a country that spreads AI’s benefits. In what he called the “San Francisco AI Declaration,” Lee said Korea would show the world how to integrate AI across industries and daily life, grounded in human-centered values. 

The vehicle would be “AI for All“: AI access for every citizen, adoption across government services, and a second growth engine—in AI models, applications, and robotics—alongside the chips of Samsung and SK Hynix.

“AI for All” is the most ambitious industrial policy bet Seoul has made in a generation. President Lee has backed a handful of domestic consortia to build a sovereign Korean model, and during the summit, the presidential office announced semiconductor supply and production partnerships between U.S. and Korean firms worth a combined USD 950 billion over five years. Korea is building AI data center capacity in the multiple-gigawatt range. But more than these investments, the centerpiece of the plan is a promise to offer general-purpose AI chatbots and public service agents, with no usage limits, to all 51 million residents in the country. Lee is betting on AI the way past presidents bet on shipbuilding, chips, and cars. If it works, Korea would join the United States and China as one of the handful of countries in control of their AI destiny, with homegrown models for a technology that will shape its economy, labor market, and national security.

Alongside this bet, Korea has also legislated ahead of its peers. The National Assembly passed the AI Basic Act in December 2024, which took full effect in January 2026, making Korea the first country in the world with a comprehensive AI law fully in force. Korea established an AI Safety Institute (AISI), and the legislation sets a compute threshold that designates high-performance systems for additional safety duties and imposes risk management obligations across ten “high-impact” domains, including energy, healthcare, hiring, public services, and nuclear operations. The European Union, whose 2024 AI Act was the first comprehensive AI law adopted worldwide, has since deferred its high-risk obligations until December 2027. The United States has no horizontal AI law at all. Measured by what is actually in force, Korea leads, and it is developing a policy approach that could demonstrate how AI prosperity and safety can be built together.

The Bet on AI

The bet on “AI for All” is about more than access to AI models and personal agents. Access is the part a government can deliver by decision—select the operators, supply the graphics processing units (GPUs), and open the service. Whether people gain from AI is decided elsewhere, in what the technology does to their work and the decisions made on their behalf.

The same ideas are being raised here in the United States. In an August 2026 essay, Meta CEO Mark Zuckerberg argued that the central risk in AI is the concentration of control, and that the answer is a capable personal agent in the hands of everyone rather than a system that automates people out of work. Korea has reached a similar conclusion by a different route and has acted on it through public policy. Both approaches rely on the assumption that providing widespread access to the technology is key to distributing the benefits of AI. With this in mind, the more the United States and Korea, including their companies, cooperate and coordinate to expand AI access, the greater the benefits will be shared across both nations. This argument only holds if AI is developed responsibly and safely, which includes world-class AI evaluation capacity and research.

Currently, the evaluation capacity of AI technologies sits with AISI. It represents the country in the International Network of AI Safety Institutes, leads a safety consortium of some two dozen industry and academic bodies, and runs a dual-use task force on chemical, biological, radiological, and nuclear risk. AISI’s mission and mandate are critical, and questions about its resourcing are now being debated. The 2027 budget, approved by the presidential cabinet on September 1, raises the Ministry of Science and ICT’s AI spending by 84 percent to USD 6.3 billion (KRW 9.4 trillion) and doubles the line for AI safety technology development to USD 10 million (KRW 15.2 billion). The National Assembly decides whether that ratio matches the program it is meant to protect. More political and financial capital invested in AISI, and closer coordination between AISI and U.S.-based safety institutes and incubators, would augment and accelerate Korea’s strong legislative move toward global leadership in safety practice.

Alongside AISI, Korea could emulate the safety and research centers now attracting significant investment elsewhere by establishing a research hub of its own, enabling Korea to develop and disseminate world-leading expertise on economic disruption, alignment, and other risks that AI creates or exacerbates. The foundations already exist. Seoul’s city government, in conjunction with some of the country’s leading universities, runs an AI hub in southeast Seoul, and Korea’s convening record, from the AI Seoul Summit and its Seoul Declaration to hosting the Responsible Artificial Intelligence in the Military Domain (REAIM) summit in 2024, is as strong as any in the region. A world-class hub would underpin Korea’s stated leadership through “AI for All” and its Group of Twenty (G20) presidency in 2028, where President Lee has already argued that AI must deliver fair opportunity to every nation.

Best case, the development of the hub is supported, guided, and financed by the Korean private sector with assistance from U.S. incubators, companies, and NGOs. 

Conclusion

In December 2026, 51 million people will get a free chatbot with no usage limits, running on Korean models. No other government has attempted anything close, and Korea will almost certainly deliver it on time. President Lee has staked his term on it. The country has an industrial base supplying the memory that the rest of the world’s AI runs on and soft power that carries weight in international policy, grounded in its relationship with the United States. Those are the foundations of a country that can champion an AI initiative that is both ambitious and equitable.

Henry Haggard is Senior Visiting Fellow at KEI. The views expressed are the author’s alone.

This material is distributed by KEI on behalf of the Korea Institute for International Economic Policy. Additional information is available at the Department of Justice, Washington, DC.