Seoul: South Korea has introduced the world’s most comprehensive set of laws regulating artificial intelligence that could become a model for other countries.
The legislation, however, has already drawn criticism from both local tech startups and civil society groups, highlighting the challenge of striking a balance between innovation and public protection.
The new AI Basic Act, which took effect last Thursday, comes amid growing global concerns over manipulated media and automated decision-making. The law requires companies offering AI services to label outputs clearly, with invisible digital watermarks for cartoons or artwork and visible labels for realistic deepfakes.
Systems classified as ‘high-impact AI,’ such as those used in medical diagnosis, hiring, or loan approvals, must undergo risk assessments and document decision-making processes. Extremely powerful AI models will need safety reports, though the threshold is so high that no systems worldwide currently meet it.
Violations can result in fines of up to 30 million won (£15,000), but the government has promised at least a one-year grace period before enforcement. Officials emphasise that 80-90 percent of the law is aimed at promoting industry rather than restricting it, supporting South Korea’s ambition to become one of the world’s top three AI powers alongside the United States and China.
Companies must determine whether their systems qualify as high-impact AI, a process critics describe as lengthy and uncertain. Concerns over competitive imbalance have also been raised, as Korean firms of all sizes must comply, while only foreign companies meeting specific thresholds, such as Google and OpenAI, are regulated.
Civil society groups argue that the legislation does not adequately protect individuals from harm caused by AI. While the law offers safeguards for users of AI systems, these are largely hospitals, financial institutions, and government agencies rather than ordinary people affected by AI outputs. Human involvement clauses create further loopholes, and the human rights commission has noted that those most at risk of rights violations remain largely unprotected.
The push for regulation comes amid serious domestic concerns. South Korea accounted for more than half of the global victims of deepfake pornography in 2023, highlighting the risks of unregulated AI content. Civil society groups continue to call for stronger measures to protect citizens.
Experts note that South Korea has taken a different path from other countries. While the European Union has adopted a strict risk-based model and the United States and United Kingdom rely on sector-specific regulations, South Korea has opted for a flexible, principles-based framework centred on ‘trust-based promotion and regulation.’ Officials say the law is designed to evolve and clarify rules through revised guidelines.
