Module 2 Knowledge Check (Self‑Check)
Use this self‑check to confirm you understand supervised use, documentation, and verification expectations.
Questions
-
Which item is most important to capture in an AI audit trail for a high-risk task?
- A. The color theme of the document
- B. The prompt, sources used, verification steps, and approver
- C. The AI’s estimated confidence score only
- D. The number of emojis in the output
Suggested answer: B.
Why: Defensibility requires traceability: inputs, instructions, checks, and accountability. -
True or False: For a court filing, it is acceptable to rely on AI-generated citations if you used an “enterprise” model.
- True
- False
Suggested answer: False.
Why: Tool settings reduce confidentiality risk, but they do not replace citation validation. -
Which verification method best matches a memo that summarizes 200 documents?
- A. No checks; trust the tool
- B. Sampling + spot checks against primary documents
- C. Grammar checking only
- D. Ask the AI to “double check itself” and stop there
Suggested answer: B.
Why: Sampling is a scalable way to validate accuracy when volume is high. -
Which statement best describes “supervised use”?
- A. AI makes final decisions so humans can focus elsewhere
- B. AI outputs are drafts reviewed, verified, and approved before they become work product
- C. AI is used only for marketing
- D. AI replaces attorney review for routine matters
Suggested answer: B.
Why: Supervision is a process and accountability model. -
When should you consider client disclosure or consent about AI use?
- A. Only when the client asks
- B. When AI use involves sharing sensitive client information with a third-party tool or materially changes risk/cost
- C. Never
- D. Only when the output is formatted as a table
Suggested answer: B.
Why: Client expectations and confidentiality risk may require communication and alignment.
If you answered fewer than 4 correctly: review the documentation checklist and write your firm’s “minimum audit trail” standard for high-risk tasks.