KIBO Introduces AI Risk-Tiered Management System, Lays Foundation for Safe AX

[Financial News] The Korea Technology Finance Corporation (KIBO) will introduce an enterprise-wide management system that classifies risks arising from the use of artificial intelligence (AI) into four tiers. It plans to expand AI use in technology assessments and customer services while ensuring safety and reliability.
 According to industry sources on the 9th, KIBO has comprehensively revised its internal regulations on AI and data. The overhaul was undertaken to prepare for the enforcement of AI-related laws and consolidate AI and data management standards that had been scattered across existing regulations and guidelines.
 The goal is to strengthen implementation of its AI transformation (AX) strategy, established late last year, and systematically manage risks that may arise during the use of AI.
 The key change is that management and approval procedures have been divided into more specific steps according to the risk level of AI systems. KIBO will classify AI systems into four tiers, from ultra-high risk to general, based on risk assessment results, and apply different approval authorities and reassessment cycles to each tier.
 Tier 1 AI systems, classified as ultra-high risk, must be approved by the AI Innovation Committee. Tier 2 systems, classified as high risk, require approval from the chief AI officer (CAIO), while Tier 3 systems require approval from the head of the department overseeing the relevant business. Before adopting and using AI, KIBO will determine whether it qualifies as high-impact AI and conduct risk and impact assessments during the development and adoption stages.
 KIBO also reorganized the body overseeing AI operations. It transformed the existing Digital Innovation Committee into the AI Innovation Committee and designated its executive director as CAIO. The aim is to strengthen enterprise-wide management capabilities by clarifying who is responsible for advancing the AX strategy and managing AI risks.
 It also strengthened its ethics and user-protection standards. KIBO established public interest, equity, transparency, accountability, safety and privacy protection as the six principles of AI ethics, and put in place a system for checks at each stage of use.
 KIBO plans to expand the use of AI in key areas, including technology assessments and customer services, and improve the effectiveness of its risk management system.
 KIBO Chairman Hyung-taek Kwon said, “AI transformation goes beyond simply adopting technology; it is a process that changes how technology assessments are conducted and how work is done across the board, making it important to ensure reliability and safety,” adding, “We will expand the use of AI based on systematic risk management and responsible operations.”
[email protected] Lee Ju-mi Reporter