引導式測驗 · 硬 等級

用於首次透過文件審查測驗的生成式人工智慧

將生成式人工智慧與技術輔助審查進行比較,包括文件工作流程的驗證、保密性和法律監督。

問題 1 的 8

How does technology-assisted review help prioritize a large document collection?

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. How does technology-assisted review help prioritize a large document collection?
    • Prioritize documents under a defined review protocol
    • Determine the outcome of a lawsuit
    • Replace all human coding
    • Create new evidence absent from the record
  2. Why should an AI-generated summary be checked against source documents?
    • Summaries always preserve every detail
    • Source documents cannot be searched
    • A fluent summary can omit a qualification or misstate content
    • Legal teams should avoid identifying relevant passages
  3. Which measure can help assess a retrieval workflow’s missed-document risk?
    • Number of pages per document
    • Precision among retrieved documents, without reviewing material that was missed
    • Average summary length
    • Number of documents processed, regardless of relevant documents not found
  4. What should a legal review team define before model-assisted review?
    • The final legal conclusion before examining documents
    • Only the preferred generative model
    • A prompt that avoids human review
    • The objective, population, privilege process, and quality checks
  5. Why preserve document identifiers and source links?
    • So findings can be traced back to specific records
    • So reviewers cannot change a coding decision
    • So a model can infer privilege automatically
    • So the original document can be discarded
  6. Which concern is especially important when processing client documents?
    • Model response speed only
    • The number of clauses in a contract
    • Confidentiality, access controls, retention, and vendor terms
    • Whether the summary uses first-person language
  7. What does one overall accuracy score fail to show?
    • Whether the system used a computer
    • Which relevant documents were missed and how review thresholds behaved
    • How many documents were in the corpus
    • Whether the prompt was readable
  8. Who should make a privilege or production decision?
    • The model based only on a confidence score
    • The vendor without matter context
    • The first reviewer who sees a summary
    • Appropriately qualified legal reviewers under the governing process