South Korea lacks a practical legal framework to respond to AI-driven security threats, with current rules built on the assumption that intrusions can be prevented, a forum heard on Thursday. Jeong Se-jin (정세진), a lawyer at Lee & Ko, said high-performance AI such as Mythos can find vulnerabilities and attack on its own, beyond existing controls. He urged regulation focused on resilience, dynamic 대응 and clearer accountability for agentic AI incidents.

South Korea\'s Ministry of Science and ICT said on May 21 it will seek public comment from that day on a draft revision to the enforcement decree for the Artificial Intelligence Framework Act, ahead of the law\'s July implementation. The draft defines the scope of AI-vulnerable groups and sets rules for prioritising AI products and services in public procurement, including liability exemptions. It also expands eligibility for cost support, specifies procedures for using a venture fund for AI startups, and sets requirements for AI research institutes.