引導式測驗 · 中等 等級

中國法院關於 AI 圖像版權的判決測驗

測試中國人工智慧圖像案件中考慮的證據和人類貢獻,以及為什麼沒有一個裁決可以解決所有輸出。

問題 1 的 8

What human contribution mattered in Beijing’s 2023 AI-image decision?

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 human contribution mattered in Beijing’s 2023 AI-image decision?
    • The user’s creative choices in prompts, parameters, refinements, and final selection
    • The model’s training compute alone
    • The platform’s subscription price
    • The image file’s dimensions
  2. Why did the later Beijing case reject the plaintiff’s evidence?
    • AI images can never be protected
    • The plaintiff used more than one prompt
    • A recreated image did not prove the original creation process
    • The image was published online
  3. Do the rulings mean every AI-generated image is copyrighted?
    • Yes, any prompt creates ownership
    • No; protection depends on originality, human contribution, and evidence in the specific case
    • No, all AI images are excluded
    • Only images made with open-source models qualify
  4. What did the 2023 court treat the AI model as?
    • The sole legal author
    • A copyright owner
    • A court-appointed expert
    • A tool used by the human creator
  5. Which record best helps show the original process?
    • A later recreation only
    • 一則產品廣告
    • Prompt history, parameter changes, drafts, selected outputs, and edits
    • A screenshot of the model’s home page
  6. What do these rulings not decide automatically?
    • All questions about training-data use and model-provider infringement
    • Whether courts apply Copyright Law
    • Whether authorship can involve human contribution
    • Whether evidence matters
  7. Why might a single judgment not create a universal rule?
    • Chinese courts cannot hear copyright cases
    • The model controls the court
    • Copyright does not apply to images
    • Different facts, evidence, and courts can lead to different outcomes
  8. What should a company do before licensing an AI image?
    • Assume the prompt alone guarantees rights
    • Review the creator’s evidence and relevant service terms separately
    • Treat model output as public domain
    • Rely only on a platform watermark