Enhancing scientific understanding of AI safety

1. 1AI risk identification and classification

   Based on global discussions, define types of risks and classify those that need to be prioritized and managed at the national level.

2. 2AI safety evaluation framework

   Develop and apply an AI safety evaluation framework (including indicators, standards, and methods) tailored to different AI risks in collaboration with domestic and international partners.

3. 3Establishment of infrastructure for AI safety evaluation

   Develop and provide datasets, tools, and resources for AI safety evaluations.

4. 4Systematic research and analysis on AI safety

   Establish a scientific analysis and knowledge base that support AI safety.

Supporting Korean AI companies in the area of AI safety

1. 1AI safety policy research

   Provide expertise in establishing AI safety policies and standards.

2. 2National security risk response

   Identifying AI-related risks from a national security perspective and devising appropriate countermeasures.

3. 3Safety consulting support

   Providing consulting services to companies on AI safety.

4. 4AI technology impact assessment

   Preemptively assess the impact of AI technology on the economy, society, employment and support the development of response policies.

5. 5Establishment of AI safety information system

   Establish a system for collecting and disseminating information about AI safety.

Advancing AI safety policies and supporting the establishment of safety systems

1. 1National collaboration hub

   Serve as a platform that facilitates bilateral or multilateral collaboration with relevant domestic organizations.

2. 2Global collaboration gateway

   Serve as the national representative in partnerships with AI safety institutes and international organizations to build global governance frameworks.