South Korea Establishes Legal Framework for AI to Combat Deepfakes and Misinformation

South Korea on Thursday formally enacted a comprehensive law governing the use of artificial intelligence, becoming the first country to introduce a legal framework specifically designed to address growing concerns about AI-related misinformation, deepfakes, and other potential risks.

The AI Basic Act establishes national standards for AI development and deployment while giving government authorities new powers to investigate misuse and enforce compliance through penalties. According to officials at the Ministry of Science and ICT, the legislation aims to balance technological innovation with necessary safeguards as AI systems become increasingly integrated into daily life.

Central to the new law is the classification of certain applications as “high-risk AI” – systems whose outputs could significantly impact individuals’ rights, safety, or economic wellbeing. This category encompasses AI used in consequential decisions such as hiring processes, credit assessments, and medical guidance.

Companies developing or providing high-risk AI services will face strict requirements, including clearly disclosing when users are interacting with artificial intelligence systems and taking responsibility for managing potential harms. Additionally, all AI-generated content must carry digital watermarks indicating its algorithmic origin.

“Watermarking is the minimum safeguard needed to prevent the misuse of AI technologies, including deepfake content,” explained a ministry official. This measure addresses growing concerns about synthetic media that can realistically mimic human appearances, voices, or writing styles, which have already created challenges for information integrity worldwide.

The law’s reach extends to major global technology companies operating in South Korea. AI service providers with annual global revenue exceeding 1 trillion won (approximately $740 million), domestic sales of at least 10 billion won ($7.4 million), or more than one million daily users in the country must appoint a local legal representative. This requirement will likely affect international tech giants such as OpenAI and Google.

Non-compliance could result in fines up to 30 million won ($22,000). However, authorities have announced a one-year grace period before full enforcement begins, giving companies time to implement the necessary changes to meet the new standards.

Beyond regulatory measures, the legislation also focuses on fostering domestic AI innovation. The science minister must present a national AI policy roadmap every three years, outlining investment priorities and long-term development strategies for the sector.

South Korea’s move comes at a time when governments worldwide are struggling to develop effective regulatory frameworks for rapidly evolving AI technologies. The European Union has been working on its own AI Act, while the United States has primarily relied on voluntary commitments from technology companies and executive orders to guide AI governance.

The timing is particularly significant as generative AI systems like ChatGPT, Google’s Gemini, and various image generators have demonstrated increasingly sophisticated capabilities, raising new questions about content authenticity, intellectual property rights, and potential social impacts.

South Korea’s technology sector has historically been dominated by hardware manufacturers like Samsung and LG, but the government has recently increased focus on software and AI development to maintain economic competitiveness. This new legislation signals the country’s intention to take a leading role in establishing governance standards for emerging technologies.

Industry analysts will be watching closely to see how South Korea’s regulatory approach affects both local and international AI development, and whether it strikes the intended balance between innovation and protection. The country’s experience could provide valuable insights for other nations considering similar legislation as AI technologies continue to advance globally.