[Insight] EU AI Act Takes Effect in 2026, Korean Companies' AI Governance Response Strategy

In August 2026, the core obligations of the EU AI Act will be fully applied. This goes beyond a mere European regulation—it is expected to function as a global standard defining transparency and accountability in global AI operations going forward, as well as a precedent for Korean-style AI regulation.

Rather than regulating technology itself, the EU AI Act adopts a Risk-based Approach that manages the Risk arising from the use of technology. It is divided into four levels according to risk level, as follows.

• Unacceptable Risk: Social credit scoring and unauthorized facial recognition systems that harm human dignity are, in principle, prohibited from being placed on the market.
• High Risk: This covers critical infrastructure such as transportation and energy, as well as areas that greatly affect individuals' opportunities, such as recruitment, education, and financial services. A Conformity Assessment and establishment of a risk management system are mandatory before market launch.
• Limited Risk: This covers areas such as chatbots and deepfakes where users must be aware that they are interacting with AI, with transparency being the key obligation.
• Minimal Risk: Low-risk areas such as spam filters, where autonomous management is permitted.