S0176RS6-T20-S0176-Z · Full risk code

Manipulation of Elections and Political Processes

選舉與政治過程操縱

Operation & Monitoring
Risk Description

When different voter groups receive divisive messages tailored to their particular concerns, mutually contradictory yet individually persuasive; due to unmitigated control gaps, because the messaging is highly dispersed and personalized, outside observers cannot grasp the full scope of the operation, triggering compliance exposure and operational reputational costs.

Framework Mappings

EU AI ActArt.50
NIST AI 600-1Information Integrity
NIST AI RMFMANAGE 2.4
ISO/IEC 42005影響評估流程
ISO/IEC 5338運作與監控
MIT AI Risk RepositoryDomain 3

Risk Treatment & Implementation Guidance

Strengthen AI-content marking and source verification during sensitive periods such as elections; Detect divergent manipulation via coordinated-behavior and cross-segment message comparison, identifying highly personalized micro-targeting; Collaborate with platforms, electoral bodies and fact-checkers to raise overall transparency