Republic of Korea: Ministry of Science and Information and Communication Technology announced drafting of guideline on AI safety obligations

Description

Ministry of Science and Information and Communication Technology announced drafting of guideline on AI safety obligations

On 15 January 2025, the Ministry of Science and Information and Communication Technology (MSICT) issued the notice on Artificial Intelligence (AI) safety obligations pursuant to Article 32 of the Framework Act on the Advancement of Artificial Intelligence and Creation of a Trust Foundation (AI Framework Act), introducing enforceable safety obligations across the lifecycle of high-performance artificial intelligence systems. High-performance artificial intelligence is defined as systems with cumulative computational capacity exceeding 10^26 floating-point operations per second (FLOPS) or with significant impacts on human life or rights. Obligations include risk identification methodologies, independent team assessments, mitigation measures, emergency response planning, stability testing, and continuous monitoring. The Notice also sets conditions for reliability certification, risk assessment indicators, and safety verification procedures.

Original source

Scope

Policy Area
Design and testing standards
Policy Instrument
Design requirement
Regulated Economic Activity
ML and AI development
Implementation Level
national
Government Branch
executive
Government Body
central government

Complete timeline of this policy change

Hide details
2025-01-15
under deliberation

On 15 January 2025, the Ministry of Science and Information and Communication Technology (MSICT) is…

2026-01-22
adopted

On 22 January 2026, the Ministry of Science and ICT released the artificial intelligence (AI) safet…