引導式測驗 · 中等 等級

圖書館員人工智慧測驗

探索圖書館如何使用人工智慧進行發現、元資料和參考工作,同時保護隱私、引用來源並讓圖書館員保持可用。

相關引導路徑圖書館員人工智慧
問題 1 的 8

圖書館探索聊天機器人應隨事實性答案提供什麼?

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. 圖書館探索聊天機器人應隨事實性答案提供什麼?
    • Links to source records and a path to librarian help
    • A confidence score as a replacement for sources
    • Only a generated paragraph
    • A claim that it searched every available work
  2. 為什麼生成的引用可能具有風險?
    • A citation in a familiar style guarantees that it supports the answer
    • A real book title proves the system verified the cited passage
    • It may be inaccurate or refer to a source that does not exist
    • An author name in a citation confirms that the source exists
  3. 什麼隱私問題適用於圖書館搜尋查詢?
    • Queries are always anonymous by default
    • They can reveal sensitive interests or personal questions
    • Search history cannot be stored by a tool
    • Only book titles can be sensitive
  4. 聊天機器人應如何回答超出其信任範圍的問題?
    • Invent a likely local policy
    • Use the newest answer from social media
    • Hide which sources it searched
    • Escalate to a librarian or clearly state the limitation
  5. AI 生成圖書館目錄詮釋資料時可能會出現什麼錯誤?
    • It may misidentify people or erase relevant cultural context
    • Metadata cannot be corrected by library staff
    • It always decreases catalog searchability
    • It guarantees consistent classification
  6. Which evaluation helps assess a discovery tool?
    • Count generated responses without checking them
    • 僅測量答案長度
    • Test accuracy, coverage, and escalation on representative queries
    • Use one staff query as a complete test
  7. How should a discovery assistant characterize search results?
    • Replace source evaluation and librarian guidance
    • Suggest candidate sources for users to inspect
    • Determine the truth of every retrieved source
    • Guarantee comprehensive coverage of all knowledge
  8. Why retain human reference support?
    • Human librarians cannot use search tools
    • Chatbots always return outdated information
    • Users should not ask follow-up questions
    • Some questions require context and judgment beyond automated retrieval