The Ministry of Science and ICT said on Wednesday that the government will provide consulting services and financial support to help companies voluntarily comply with the new obligations rather than imposing immediate fines, a stark contrast to the EU AI Act which began enforcing bans on certain AI systems from February 2025.
The law, officially titled the Framework Act on the Development of Artificial Intelligence and the Creation of a Foundation for Trust, marks a shift from non-binding guidelines to a legally enforceable framework. It establishes grounds for research and development support, training data infrastructure, and workforce development while mandating transparency and safety measures.
Under the new regulation, generative AI providers must attach watermarks or identification labels to AI-generated content to combat deepfakes and misinformation. Operators of "high-impacting AI" systems that could significantly affect lives, safety, or fundamental rights will face heightened risk management requirements.
Separately, the ministry announced that major platform operators including Naver, Kakao, and Netflix will be required to strengthen customer service systems from Feb. 12, 2026. These companies must provide multi-channel support including AI chatbots and automated response systems, with real-time human agent connections when needed.
The government also unveiled plans to launch the K-STAR visa program from January to attract top science and technology talent from overseas. Foreign students in science and engineering fields who receive university presidential recommendations will be eligible to apply for residency or permanent resident status regardless of employment status.
South Korea's measured approach mirrors Japan, which enacted its AI Promotion Act in May 2025 with no explicit penalties, relying instead on voluntary compliance and reputational incentives. The United States remains without comprehensive federal AI legislation, with U.S. President Donald Trump signing an executive order in December seeking to preempt state-level regulations while calling for a "minimally burdensome" national standard.
China has quietly removed plans for a comprehensive AI law from its 2025 legislative agenda, opting instead for a phased approach that prioritizes pilot programs, technical standards, and targeted rules over early codification. Beijing is relying on existing statutes and sector-specific measures to manage AI-related risks while keeping compliance costs low, though observers note this leaves firms navigating fragmented frameworks and overlapping obligations.
Copyright ⓒ Aju Press All rights reserved.
