South Korea became one of the first countries to implement a comprehensive law regulating artificial intelligence. According to the new rules, materials created using artificial intelligence - pictures, videos, audio, and dipfeiks - must be marked with a special watermark. That is, the user must clearly know that the content is a "SI product."
These requirements are established within the framework of the "Basic Law on Artificial Intelligence" (AI Basic Act), which entered into force in January 2026. The main idea of the law is simple: AI services should be transparent, and society should not fall prey to invisible manipulation.
What content will be marked?
The rules apply to materials created or edited by a generative AI. Especially in the case of dipfake, control is being strengthened: if someone makes a voice, face, or video look "real," this should no longer remain hidden. The "watermark" requirement is mentioned as a marking mechanism.
In "high-risk" industries, the requirements are stricter
The new law pays special attention to areas considered high-risk. Including:
- healthcare,
- transport,
- finance,
- safety
The mechanism of human control and responsibility comes to the forefront in the use of artificial intelligence in such areas.
Responsibility and "transition period"
It is noted that administrative measures (fines) are provided for violators of the rules. At the same time, the government said that a transition period will be given for companies to adapt to the new requirements - initially, there is a position of "not overrun" with sanctions.
Why is this step important?
In short: the question of "who is right and who is wrong" on the internet is becoming more and more difficult day by day. AI-generated video and audio have become a powerful tool for persuading people. South Korea, however, recognized this risk early and began to establish legal rules - the goal is to reduce disinformation and abuse, thereby maintaining public trust.
The conclusion is this: now the period of receiving content not in "saw - believed" mode, but in "saw - verified source" mode is accelerating even more. We live with AI, but we also unveil its "mask" - this is a trend.
Read “Zamin” on Telegram!
