AI Law Delays Spark Concerns in South Korea

Uncertainty Rises as AI Law Decrees Miss Schedule

South Korea has missed its self-imposed deadline of June 30 to disclose draft versions of enforcement decrees and rules for its basic artificial intelligence law. The National Assembly overwhelmingly passed the AI Basic Act — officially known as the Framework Act on the Development of Artificial Intelligence and Establishment of Foundation for Trust — in late December 2024. This legislative move took markets by surprise, especially as the country was undergoing significant political turmoil following the failed attempt by the then-President Yoon Suk Yeol to impose martial law, which led to widespread unrest.

The subsequent promulgation of the law by the Cabinet Council in January came unexpectedly, as the government was operating under an acting president. The Ministry of Science and ICT stated at that time that it would release draft versions of the enforcement decrees and rules by the end of June for public comment, aiming for a finalized version early enough for companies and users to prepare for the January 2026 implementation of the law. The goal was to ensure smoother compliance for stakeholders.

Political Context and Delays

On the surface, the delay seems understandable. President Lee Jae Myung, who had only been in office for just over a month, has yet to fully stabilize his administration. Lee assumed office without the usual transition period due to the election held to fill the presidency vacated by the forced removal of Yoon by the Constitutional Court.

However, companies are becoming increasingly anxious as the uncertainty surrounding the subordinate legislation process prolongs. They must strategize for what is anticipated to be one of the region’s major markets for AI business opportunities, all while operating with limited clarity and time.

International Regulatory Landscape

The delay in South Korea’s AI legislation arrives as many stakeholders express concerns that, while enacting the law later than the European Union, South Korea might be the first jurisdiction globally to implement regulatory obligations. The EU is not expected to enforce its provisions until several months after South Korea does.

For example, while South Korea’s AI Basic Law is set to take effect in late January 2026, most regulatory provisions of the EU AI Act will commence from August 2026, with staggered deadlines. This creates a scenario where South Korean companies may face regulatory hurdles before their European counterparts.

Shifting Global Dynamics

When South Korea’s government and National Assembly rapidly passed the law in late 2024, it was believed that having a regulatory framework would benefit the nation’s AI firms in major markets, where strict regulations were anticipated. However, the global AI policy landscape has shifted dramatically, especially following announcements from companies like DeepSeek and other Chinese developers regarding their successful development of world-class AI systems.

Countries such as the United States, Britain, Canada, and Germany have recently adopted bold measures to promote AI development, with some even moving away from stringent regulations. This global trend raises concerns that South Korea may have rushed into a regulatory-first approach, potentially hindering innovation.

Proposed Legislative Revisions

In response to these uncertainties, the then-opposition Democratic Party of Korea — now the ruling party — proposed a revision bill in April to suspend regulatory provisions for three years. Lawmakers emphasized that AI-related trends in the U.S., EU, and Japan are moving away from regulation towards promotion and industry growth.

There is currently no indication that this revision bill will progress quickly, as the leader of the proposing lawmakers acknowledged that approval may take time. This situation has prompted the Ministry of Science and ICT to explore options to effectively delay the full implementation of the bill’s obligations to avoid stifling innovation.

According to media reports, the ministry suggested introducing a grace period on the provision imposing fines for violations listed in the law to create an effect similar to delaying regulatory measures. This flexibility could be essential for companies aiming to adapt to the evolving landscape.

Unresolved Issues and Future Outlook

All matters related to subordinate legislation remain highly uncertain, as President Lee has yet to reveal how he intends to implement his election pledges. This includes a pledge to revamp the Presidential Committee on AI, which was established in late 2024 but has been inactive for months due to political unrest.

Experts express concerns over how many of the plans prepared by the outgoing government will be upheld by the new administration and how quickly South Korea can resolve the ongoing uncertainties surrounding conflicting regulatory obligations.

As uncertainty looms, one consensus emerges: uncertainty is detrimental to both the business environment and investor sentiment towards South Korea, particularly in the rapidly evolving sector of artificial intelligence.

More Insights

Critical Evaluations of AI Compliance Under the EU Act

The EU’s Artificial Intelligence Act introduces new obligations for organizations regarding general-purpose AI models, set to take effect in August. Dealmakers must enhance their due diligence...

Microsoft’s Science Chief Opposes Trump’s AI Regulation Ban

Microsoft's chief scientist, Dr. Eric Horvitz, has criticized Donald Trump's proposal to ban state-level AI regulations, arguing that it could hinder progress in AI development. He emphasizes the need...

AI Regulation: Europe’s Urgent Challenge Amid US Pressure

Michael McNamara discusses the complexities surrounding the regulation of AI in Europe, particularly in light of US pressure and the challenges of balancing innovation with the protection of creative...

Decoding the Regulation of Health AI Tools

A new report from the Bipartisan Policy Center examines the complex regulatory landscape for health AI tools that operate outside the jurisdiction of the FDA. As AI becomes more integrated into...

Texas Takes the Lead: New AI Governance Law Unveiled

The Texas Responsible Artificial Intelligence Governance Act (TRAIGA), passed on May 31, 2025, establishes disclosure requirements for AI developers and deployers while outlining prohibited uses of AI...

Texas Enacts Groundbreaking AI Governance Law

On June 22, 2025, Texas Governor Greg Abbott signed the Texas Responsible AI Governance Act (TRAIGA) into law, significantly altering the original draft that proposed strict regulations on "high-risk"...

G7 Summit Fails to Address Urgent AI Governance Needs

At the recent G7 summit in Canada, discussions primarily focused on economic opportunities related to AI, while governance issues for AI systems were notably overlooked. This shift towards...

Africa’s Bold Move Towards Sovereign AI Governance

At the Internet Governance Forum (IGF) 2025 in Oslo, African leaders called for urgent action to develop sovereign and ethical AI systems tailored to local needs, emphasizing the necessity for...

Top 10 Compliance Challenges in AI Regulations

As AI technology advances, the challenge of establishing effective regulations becomes increasingly complex, with different countries adopting varying approaches. This regulatory divergence poses...