Answer all 12 scored questions, then submit. Several questions (8, 9, 10) draw directly on the skin-lesion triage tool proposal from Handout Section E — if you haven't worked through it yet, do that first. Question 13 is a short reflection and isn't scored.
Question 1
Per Section A, does a patient's ordinary consent to clinical care also count as consent for a specific AI tool to process their data?
Correct answer: the second option — ordinary clinical consent doesn't automatically extend to a specific third-party AI tool's data handling.
Question 2
Why does it matter where a cloud-hosted AI tool's servers are physically located?
Correct answer: the first option — server location determines which jurisdiction's data-protection rules, if any, apply.
Question 3
Per Section B, what is the current state of AI-specific regulatory oversight for clinical software in Timor-Leste?
Correct answer: the third option — this is the honest starting fact Section B works from.
Question 4
Does the absence of an AI-specific regulatory framework reduce a physician's accountability for a decision made with AI assistance?
Correct answer: the second option — Section D is explicit that a governance gap doesn't reduce responsibility.
Question 5
Per Section D, what is a physician's responsibility beyond individual clinical decisions, in a system with a governance gap?
Correct answer: the first option — professional responsibility runs in two directions: careful individual judgment and constructive advocacy.
Question 6
What is the purpose of the Adoption Checklist introduced in Section C?
Correct answer: the third option — it's built as a FIND-category tool, meant to be looked up and run through, not memorized.
Question 7
Which checklist item specifically addresses whether a clinician can meaningfully overrule a tool's output during a real, time-pressured encounter?
Correct answer: the second option — a theoretical override that isn't usable in practice isn't a real override.
Question 8 · Applied — from the worked exercise
In the Handout's Section E exercise, a regional health office proposes a smartphone-based AI skin-lesion triage tool, validated in Southeast Asia but not specifically in Timor-Leste, to extend triage capacity to rural health posts.
What is the clearest concern raised by the tool's validation population?
Correct answer: the first option — the discussion treats this as a real, nameable gap rather than either dismissing it or treating it as disqualifying on its own.
Question 9 · Applied — from the worked exercise
Why does it matter that the tool would be used by non-specialist staff at rural health posts rather than by a dermatologist?
Correct answer: the third option — a theoretical override that assumes specialist judgment isn't a real override for this specific user group.
Question 10 · Applied — from the worked exercise
What was the discussion's overall verdict on whether the health office should adopt the tool?
Correct answer: the second option — deliberately, per the discussion's closing line, so the decision-maker has real information rather than a predetermined answer.
Question 11
Per Section F, what is the recommended way to raise a validation concern with institutional leadership?
Correct answer: the first option — a blanket refusal forecloses the conversation, where a pilot request opens it.
Question 12
When a vendor or ministry contact gives a vague or absent answer about a tool's validation population, what should that be treated as?
Correct answer: the third option — Section F treats the absence of a clear answer as data, not a dead end.
Question 13 · Reflection (not scored)
In your own words: think of one AI tool already in use, or being considered, in your own setting. Run it briefly through two or three items from the Adoption Checklist — what do you already know, and what would you need to find out?