Korea's AI law requires watermarks on generated content (koreajoongangdaily.joins.com)

Korea has enacted the world’s first comprehensive AI law, mandating that content generated by AI—including images, videos, and audio—must carry watermarks for transparency. This legislation, known as the AI Basic Act, aims to combat misinformation and promote accountability within the burgeoning AI sector. However, enforcement challenges loom as many watermark-removal tools are readily available, and content created by foreign applications largely evades regulation. Only businesses are required to comply with these rules, which could place domestic companies at a disadvantage compared to their less-regulated international counterparts. The law indicates a significant move towards regulating generative AI, pushing for clear labeling on media such as deepfakes, which necessitate visible watermarks. While the government has committed to a grace period without fines as the law is implemented, there remain concerns about potential reverse discrimination against Korean AI firms, who may face stricter rules than foreign companies. The law does not adequately address the proliferation of low-quality AI-generated content, also known as “AI slop,” raising further questions about its overall effectiveness in maintaining content quality and trustworthiness in the AI landscape.