Key Requirements Under South Korea’s AI Law
South Korea is leading in AI regulation and has ambitions to becomes world’s top three AI powerhouses. The South Korean government says that the new Artificial Intelligence framework is designed to support companies. But AI startups say that the new law is vague and brings a lot of compliance burdens. Under the new AI Basic Act, human oversight is necessary for “high-impact” AI applications in areas such as healthcare, nuclear safety, drinking water production, transport, loan screening, and credit evaluation. In these areas, companies must disclose in advance to users that Generative AI was used. For AI-generated content, companies have to clearly label so that users can distinguish between what’s real and what’s AI-generated output. South Korea’s Ministry of Science and ICT said that the law is designed to balance AI adoption with safety and public trust in mind. Companies will have a grace period of one year to implement these changes. If a company fails to do so, fines will be imposed. And penalties can reach up to $20,400 for failing to label Generative AI content. In light of this, AI startups have raised their concern saying that the law is making firms overly cautious, and may hinder innovation. The South Korean government is considering to extend the grace period of one year. The US, on the other hand, is not going for heavy AI regulation to avoid slowing innovation. The EU, similar to South Korea, has taken a strict route and proposing heavy penalties for compliance failures. Name Email ID
Δ
