S0150Critical

AI 生成兒少有害內容

Operation & Monitoring

Risk Description

AI systems are exploited to generate harmful content targeting minors, including but not limited to inappropriate guidance content, bullying materials, and content potentially negatively impacting childrens psychological development. AI chatbots may provide age-inappropriate advice or information when interacting with underage users. Minor users may not be properly identified and protected, being exposed to AI interaction content that should be restricted.

Framework Mappings

iso 23894R8
tw principle永續發展與福祉
tw principle人類自主
tw risk type社會系統性影響(Societal Systemic)

Controls English translation pending

  • 年齡識別機制
  • 兒少保護護欄

Implementation Steps English translation pending

  1. 實施使用者年齡識別機制,為兒少使用者提供增強保護
  2. 對兒少可能接觸的 AI 服務設立額外的內容安全護欄
  3. 提供家長控制工具,允許監護人設定互動限制
  4. 與兒少保護組織合作制定 AI 兒少安全標準