South Korea enacts AI Basic Act as a national framework including mental health provisions

South Korea passed the AI Basic Act on January 22, 2026, creating a country-wide legal framework that addresses generative AI, deepfakes, labeling and mental health in broadly stated terms. The law sets up a National AI Committee, uses a single "High-Impact AI" category, and contains some enforcement provisions while leaving many details open.

South Korea enacted the Basic Act on the Development of Artificial Intelligence and the Establishment of a Foundation for Trustworthiness on January 22, 2026. The statute is presented as a comprehensive, country‑wide regulatory framework aimed at governing modern AI systems, including generative AI and LLMs. The law creates new institutional roles and review processes while touching on risks such as deepfakes, misinformation and potential harms to mental health. Coverage of mental health impacts is included in broad language but the provisions are described in the source as modest and not tightly specified. Main provisions and context: - Enactment date: January 22, 2026, under the formal title Basic Act on the Development of Artificial Intelligence and the Establishment of a Foundation for Trustworthiness. - Institutional setup: establishes a National AI Committee to oversee implementation and interpretation; the law will be reviewed and renewed every three years. - Scope and categorization: the law distinguishes a single "High-Impact AI" category rather than multiple tiers; the definition of that category is described as complex and likely to prompt legal debate. - Content and articles: the statute contains 43 Articles plus a three-Article addendum; most provisions enter into force one year after promulgation, while a clause on digital medical devices takes effect January 24, 2026. - Enforcement and transparency: includes an obligation to label AI outputs (Article 31) and an administrative fine provision (Article 43) with penalties up to 30 million Korean won; the application of labeling rules and fines is described as unclear in practice. - Mental health provisions: Article 27 lists ethical principles that include preventing harm to mental health, but the source characterizes these safeguards as sparse and vague compared with some U.S. state laws such as those in Illinois, Utah and Nevada. Summary: South Korea's AI Basic Act establishes a national regulatory framework that addresses generative AI, transparency, and potential harms including mental health impacts, while creating a National AI Committee and a three‑year review cycle. The law contains enforcement clauses and a timetable for coming into force, but many substantive details and definitions remain open to interpretation and further rulemaking.