Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ
Bài kiểm tra về quyền riêng tư của người bảo trợ thư viện và AI
Kiểm tra quyền riêng tư của người bảo trợ thư viện: truy vấn và mượn dữ liệu, giữ chân nhà cung cấp, tiết lộ AI, chọn không tham gia, hoạt động của nhân viên, cá nhân hóa và quyền truy cập không phải AI.
Câu hỏi 1 của 8
A public chatbot asks a librarian to paste a patron’s named borrowing history. What is the safest response?
Xem trước mọi câu hỏi8 câu hỏi
Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.
- A public chatbot asks a librarian to paste a patron’s named borrowing history. What is the safest response?
- Do not disclose it to an unapproved tool; follow library privacy policy and use a data-minimized alternative
- Paste it because the tool is convenient
- Remove the patron’s first name but keep the full history
- Use the history to train all staff prompts
- What does ALA guidance recommend when third-party AI processes patron data?
- Hide the feature because the vendor controls it
- Disclose the practice, explain choices, and provide a non-AI path when feasible
- Assume patrons have consented by using the catalog
- Store all searches permanently
- Why should a library ask whether a vendor uses prompts for model training?
- Training use makes every response accurate
- Prompts never contain personal context
- Prompts may contain private searches or reference questions that could be reused
- It determines whether books can be cataloged
- Which feature preserves discovery for a patron who declines personalization?
- A personalized carousel tied to borrowing history
- A search page that hides its ranking logic
- A tool that requires account login for every query
- Known-item search or subject browsing that does not require an AI profile
- Why can de-identifying a name be insufficient?
- Search topics, timestamps, or joined records can still reveal who a patron is
- A one-way hash of an account ID always prevents linkage to other records
- Reading activity is sensitive only when a patron’s full name appears
- Search queries cannot reveal interests unless a book is borrowed
- Which vendor question is important before enabling an AI discovery feature?
- Whether the vendor can disable the feature but not how prompts are retained
- What data it collects, retains, shares, trains on, and how deletion works
- Whether searches are linked to accounts, but not what downstream services receive
- Whether data are encrypted at rest, without asking how they are used or deleted
- How can personalized ranking affect intellectual freedom?
- It guarantees equal visibility for every book
- It removes the need for subject browsing
- It may narrow or alter the material a patron sees based on inferred interests
- It prevents libraries from tracking searches
- What should staff do when a vendor cannot explain how patron prompts are retained?
- Assume the prompts are deleted immediately
- Send more data to test the service
- Let each employee choose whether to disclose patron details
- Limit, defer, or decline the feature until the library can assess the risk