A new law establishes safety, transparency, and oversight obligations for the development and use of AI.
South Korea brought into force its Basic Act on Artificial Intelligence on January 22, 2026, a broad legal framework regulating the development, deployment, and oversight of artificial intelligence in the country. The law sets requirements for human oversight of high-impact AI systems, mandates clear labeling for generative AI outputs, and applies to companies operating in or entering the South Korean AI market.
Under the law, operators of “high-impact” AI systems, including uses in healthcare, transportation, nuclear safety, and financial services, must ensure human supervision of the systems’ outputs. Products and services powered by generative or high-impact AI must provide advance notice to users and be labeled clearly when AI-generated content cannot be easily distinguished from human-created material.
The Act also establishes obligations for transparency, safety, and operator responsibilities, and includes provisions to secure AI expertise and support small and medium enterprises and startups.
The law took effect immediately upon enactment, and entities subject to its provisions are granted at least a one-year grace period before administrative penalties are enforced. Fines for non-compliance can reach up to 30 million won (about $20,400) for failures such as improper labeling of AI-generated content; subordinate regulations and detailed enforcement guidance are expected to be issued by the Ministry of Science and Information and Communications Technology.
“The legal framework was designed to promote AI adoption while building a foundation of safety and trust,” a spokeswoman for the Ministry of Science and ICT said in a written statement.
South Korean technology startups have expressed concern that aspects of the law’s language remain unclear, and that compliance could pose burdens for smaller firms developing innovative AI applications. Lim Jung-wook, co-head of South Korea’s Startup Alliance, said there was “a bit of resentment — why do we have to be the first to do this?” in reference to being an early mover in AI regulation.
President Lee Jae-myung acknowledged industry concerns, urging policymakers to support ventures and adapt measures to minimize burdens during implementation.