An artificial intelligence drawing robot on Oct. 29, 2025, creates the portrait of a visitor at K-Business Square of the APEC Economic Exhibition Hall in Gyeongju, Gyeongsangbuk-do Province. (Park Dae Jin)

By Margareth Theresia

The Basic Act on the Development of Artificial Intelligence (AI) and Creation of a Foundation for Trust (AI Basic Act) took effect on Jan. 22.

This is the world's first comprehensive legislation governing AI that covers everything about the sector, not just AI-related parts.

As a comprehensive law on AI, the act codifies national AI policy and a corresponding administrative system. The goal is to systematically develop the industry and set the basis for its safe use.

To minimize confusion in the field, provisions on transparency are subject to a grace period of at least a year, thus fact finding and fines will be suspended during this period.

The act supports the stimulation and innovation of AI and sets systematic mechanisms to ensure safety and trust. It also stipulates the legal basis for setting up a command and governance system to oversee and coordinate national AI policy.

Also specified is a range of policy measures to promote AI: support for research and development; creation of learning data; backing for AI adoption and use; promotion of startups; expansion of AI convergence; training of AI personnel; and assistance in the construction of AI data centers. Detailed standards and procedures are covered in the enforcement ordinance.

In setting the basis for safety and trust, higher legal effectiveness came through stipulating AI ethics, verification and certification systems to secure transparency and safety, as well as a management structure for high-impact AI, which is considered to pose major risks or impacts on human life, physical safety or fundamental rights.

The effectuation of the act raises systemic user protection by ensuring AI transparency, the Ministry of Science and ICT said. The requirement of prior notification allows users to know in advance the application of high-impact or generative AI when using AI products or service. 

The government set up differentiated criteria based on technical and practical considerations that distinguish between AI-generated outcomes used solely in a service environment and those transferred externally. 

"Applying watermarks to AI-generated works is a minimal safeguard against technological side effects such as misuse of AI-manipulated images, and this is a universal trend that global companies already use," the ministry said. "We hope that the effectuation of this act is an opportunity to remove legal uncertainty and create a healthy and safe domestic AI ecosystem."

margareth@korea.kr