引導式測驗 · 中等 等級

WHO關於健康AI的指導測驗

測試世界衛生組織的道德原則、自主權、問責制、公平性、生命週期治理以及與產品認證的差異。

問題 1 的 8

What does WHO’s AI-in-health guidance provide?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What does WHO’s AI-in-health guidance provide?
    • A risk-based checklist that substitutes for national health law
    • Ethical and governance principles for responsible AI use
    • A binding regulation adopted by every WHO member state
    • A product-certification scheme for vendors
  2. What does protecting human autonomy mean in WHO’s framework?
    • Autonomy applies to data consent but not decisions about care
    • A clinician’s approval alone satisfies autonomy even without patient involvement
    • People remain in control of health systems and medical decisions
    • Patients must override any clinical recommendation in every circumstance
  3. Why establish a redress mechanism?
    • People need a way to question and seek a response to harmful algorithmic decisions
    • A complaint form with no review or response pathway
    • An updated score without explanation or remedy
    • An operator may close complaints without examining them
  4. How should inclusion and equity affect AI governance?
    • Assess who benefits, who may be excluded, and access across groups
    • Use the same aggregate score for every group without examining access
    • Prioritize availability to well-resourced hospitals because they produce more data
    • Treat equal model outputs as proof that access is equitable
  5. Is WHO guidance a product certification?
    • Yes; WHO guidance replaces local approval whenever a tool is used clinically
    • No; it is a governance framework and does not replace local regulation
    • Yes; publication by WHO certifies a vendor’s product
    • No; WHO guidance applies only to medical-device manufacturers
  6. When should an organization revisit governance controls?
    • Only before initial deployment because later review invalidates evaluation
    • When data or a model changes, but not when clinical workflow changes
    • Only after a regulator orders review
    • When models, data, or workflows change
  7. Which disclosure most directly explains a health AI tool’s intended use, evidence, and limitations?
    • Accessible documentation of intended use, evidence, and limitations
    • An aggregate accuracy score without use conditions or limitations
    • Source code without evidence or intended-use information
    • An internal report unavailable to health workers and affected users
  8. Who shares responsibility for ethical AI in health?
    • Health ministries and vendors, excluding clinicians and communities
    • Developers and regulators, excluding care providers and patients
    • Developers, providers, governments, and affected people
    • Patients and vendors alone, without health systems or public agencies