South Korea has launched a landmark set of laws to regulate AI before any other country or bloc (the EU’s regulations are set to go into effect in stages through next year).
Under Korea’s AI Basic Act, companies must ensure there is human oversight for “high-impact” AI in fields like nuclear safety, drinking water, transport, healthcare, and financial uses like credit evaluation and loan screening.
Companies must also give users advance notice about products or services using high-impact or generative AI, and clearly label AI-generated output that is difficult to distinguish from reality.
Companies will be given a grace period of at least one year before fines start coming. The penalties can reach up to KRW 30 million for failing to label generative AI (that’s about $20,480).
Startups in the country are worried about the law’s vague provisions, which could lead to safe but less innovative approaches to avoid regulatory risk. The Ministry of Science and ICT will provide a guidance platform and a dedicated support center during the grace period.
