S0067RS2-T05-S0067-Z · Full risk code

Failure of Ethical Judgment

倫理判斷失當

Verification & Validation
Risk Description

When an organization uses a system to assist with cases involving value trade-offs; due to unmitigated control gaps, the system offers a single, definitive recommendation on a highly contested situation, and after staff adopt it, the affected parties react strongly and external criticism follows, triggering compliance exposure and operational reputational costs.

Framework Mappings

NIST AI 600-1Human-AI Configuration
OWASP Top 10 for LLMLLM09
NIST AI RMFGOVERN 3.2
ISO/IEC 5338驗證與確效
MIT AI Risk RepositoryDomain 5

Risk Treatment & Implementation Guidance

Prohibit single definitive recommendations on highly value-laden cases, routing them to human experts; Mark ethical boundary scenarios in system design, surfacing plural perspectives and uncertainty when touched; Establish expert review and collective decision procedures for contested cases so no single officer adopts the system's advice alone