[Background]

o Following the "Framework Act on Artificial Intelligence," safety/reliability measures for high-impact AI operators were established.

o Detailed management plans are needed to build a trust-based AI ecosystem.

[Main Points]

o Establishment of risk management policies and organizational systems.

o Compliance throughout the AI lifecycle and management of change history.

o User protection measures, including data management and preventing adversarial attacks.