South Korea has become the first country to fully implement laws regulating artificial intelligence systems. The new regulations require companies to label AI-generated content, including deepfakes, and ensure human oversight of computer systems in critical areas, from nuclear safety to healthcare.

The country's authorities claim the legislation is aimed at strengthening public trust in artificial intelligence technologies and improving the safety of their use. The South Korean Ministry of Science and Information Technology called the set of regulations the second-best in the world, after the European one.

The key difference between the Korean approach and the European one is the immediate entry into force of the entire package of regulations. The EU AI laws, approved by the European Parliament in June 2024, are being implemented in stages and will not be fully implemented until 2027.

According to the new regulations, companies are required to inform users about the use of highly effective or generative AI in their products and services. Special requirements are established for high-risk sectors: nuclear energy, water supply, transportation, medicine, and financial services. In these areas, humans must retain control over algorithmic decisions.

The global community continues to seek a balance in AI regulation. The United States has opted for a more liberal model, fearing it could stifle innovation. China, meanwhile, has already introduced its own restrictions and proposed creating an international body to coordinate the regulation of AI technologies.