引導式測驗 · 中等 等級

幽靈勞動:AI 背後的隱形勞力測驗

認識人工智慧產品背後的人工任務、它們在生命週期中發生的位置,以及為什麼勞動條件屬於供應鏈審查。

問題 1 的 8

Which task is an example of ghost work?

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. Which task is an example of ghost work?
    • Changing a company logo
    • Installing a keyboard
    • Writing a customer invoice
    • Labeling images for an object-recognition model
  2. When can human data work occur in the AI lifecycle?
    • Only before model training
    • During data preparation, training feedback, evaluation, and post-launch review
    • Only after an AI product is discontinued
    • Only during advertising
  3. Does describing work as “ghost work” mean every worker is anonymous?
    • No; it highlights that the labor is often invisible to users, not that every worker is anonymous
    • Yes, all records are hidden
    • Only for image labeling
    • Yes, because platforms erase names
  4. Why should wage claims be tied to a specific study or reporting period?
    • All workers earn the same wage
    • Wages are unrelated to work conditions
    • Pay varies by task, location, contract, and time
    • Platforms do not pay workers
  5. Why are annotation labels part of data provenance?
    • They reflect the work and judgments of the people who created them
    • They are always objective facts
    • They contain no human decisions
    • They eliminate the need for evaluation
  6. What should responsible supplier due diligence include?
    • Only model accuracy
    • Only source-code access
    • Only office location
    • Labor, safety, and grievance information as well as performance
  7. Why retain annotator expertise and provenance with labeled data?
    • It proves every label is objective
    • It makes the worker anonymous
    • It records how task instructions and human judgments shaped labels
    • 它消除了審查的需要
  8. Which statement about automation and labor is supported?
    • Automation always eliminates human work
    • AI systems often depend on human labor across their lifecycle
    • Human labor occurs only in manufacturing robots
    • All annotation is performed by AI