AI regulation in South Korea (2026)

South Korea's Framework Act on the Development of Artificial Intelligence and Establishment of Trust — along with Presidential Decree No. 36053 — took effect on 22 January 2026, making Korea the second jurisdiction after the EU with a comprehensive AI law in force. It imposes risk-tiered obligations on 'high-impact' AI providers (compute threshold ≥10²⁶ FLOPs for frontier models), mandates generative-AI output labeling, requires AI risk assessments, and obliges foreign companies to designate a local representative.

During the transition year the Ministry of Science and ICT published a legislative notice for the Enforcement Decree, specifying that AI systems trained with cumulative compute of at least 10²⁶ FLOPs are classified as 'high-performance AI' subject to enhanced safety obligations. MSIT also opened a legal-support desk to help companies prepare, and committed to a minimum one-year grace period before imposing administrative fines.

South Korea's acting government promulgated the AI Basic Act on 21 January 2025, starting a one-year transition period before enforcement. The law consolidated 19 competing AI governance bills and made Korea the second jurisdiction worldwide, after the EU, to enact comprehensive AI legislation.

Following a national crisis over AI-generated sexual abuse imagery—including cases predominantly involving minors—the National Assembly amended the Act on Special Cases Relating to Punishments for Sexual Offences. The amendment raised the maximum sentence for producing or distributing synthetic sexual content from five to seven years, and newly criminalized the mere viewing or possession of such material (up to three years or ₩30 million fine).

The Personal Information Protection Commission published its Guideline on Processing Publicly Available Data for AI Development and Services, clarifying when internet-accessible personal data may lawfully be used to train AI models under PIPA's 'legitimate interest' basis, and prescribing required safeguards, data-minimization standards, and data-subject rights mechanisms.

The National Assembly amended the Public Official Election Act (Article 82-8) to prohibit creating or distributing AI-generated synthetic audio, images, or video intended to influence elections within 90 days of polling day. The National Election Commission reported catching 129 violations in the first three weeks of enforcement during the January–February 2024 monitoring window ahead of the April 2024 general elections.

The Personal Information Protection Commission published its 'Policy Direction for Safe Use of Personal Information in the Age of Artificial Intelligence,' establishing principles for applying PIPA at every stage of AI development and deployment, and announcing the creation of a dedicated AI Privacy Team. This was the first formal regulatory guidance extending South Korea's privacy law to AI systems and signaled the PIPC's intent to act as a front-line AI regulator.

South Korea became the first country in the Asia-Pacific region to introduce comprehensive AI legislation with the Bill on Fostering Artificial Intelligence and Creating a Foundation of Trust. The bill did not pass before the parliamentary term ended, but it established the foundational 'high-impact AI' risk classification and sector-specific obligation template that carried through four years of subsequent drafts before final passage in 2024.

The Ministry of Science and ICT published South Korea's first National AI Strategy, setting a 2030 vision of becoming a top-3 AI nation. The strategy organized 9 pillars and 100 government-wide action tasks spanning AI ecosystem infrastructure, sector utilization, and people-centered AI ethics, and initiated the regulatory sandbox approach that allowed AI experimentation without full legislative constraint.

Within days of Google DeepMind's AlphaGo defeating Korean Go champion Lee Se-dol in March 2016, the government announced an emergency 1 trillion won (~$860 million) AI investment fund. This watershed moment launched South Korea's formal AI policy era, directly causing creation of new national AI research institutions, mid-to-long-term development roadmaps, and the eventual 2019 National AI Strategy.