Answer all 12 scored questions, then submit. Several questions (5, 6, 7, 10) draw directly on the Dili fever case from Handout Section D — if you haven't worked through it yet, do that first. Question 13 is a short reflection and isn't scored.
Question 1
The central idea of this module — an AI decision-support output should be treated as —
Correct answer: one input among several — the "second opinion, not a verdict" framing from Handout Section A.
Question 2
Why does this module raise the stakes compared to Module 2's documentation tools?
Correct answer: the third option — the deliberate reason decision support was sequenced after documentation in this course.
Question 3
A differential-diagnosis support tool given a vague, incomplete symptom description will typically —
Correct answer: it answers with what it has, confidently, without signaling the gap — the "garbage in, garbage out, quietly" limit from Handout Section B.
Question 4
A chest X-ray computer-aided detection tool flags nothing abnormal. What is the safest interpretation?
Correct answer: the second option — the "silent false negative" failure mode from Handout Section C.
Question 5 · Applied — from the case exercise
In the Handout's case, a febrile patient with recent floodwater exposure in Dili received an AI-generated differential ranking "viral syndrome, self-limiting" first, with dengue ranked low and leptospirosis absent entirely.
What does this most clearly demonstrate?
Correct answer: the third option — the validation-gap concept from Handout Section E, illustrated directly by this case.
Question 6 · Applied — from the case exercise
In that same case, why does the recent floodwater exposure matter clinically, independent of anything the AI tool output?
Correct answer: the exposure history is a specific, locally-relevant clue for leptospirosis — exactly the kind of context the AI tool never saw and wasn't weighted to prioritize.
Question 7 · Applied — from the case exercise
According to the case discussion, what should have determined the clinician's final working differential?
Correct answer: the second option — local epidemiological reasoning, formed independently, drives the decision; the AI ranking is one input, and in this case a misleading one.
Question 8
Why does this handout insist on forming your own differential before looking at the AI-generated one?
Correct answer: the third option — you can't un-anchor yourself once you've seen a confident-sounding ranked list.
Question 9
Per Module 1, the moment of highest risk when using an AI decision-support tool is —
Correct answer: agreement with your own gut is when scrutiny naturally drops — the confirmation-bias trap carried forward from Module 1 and sharpened in this module.
Question 10 · Applied
A colleague says: "The AI flagged something on this chest X-ray with a 94% confidence score, so I'm confident it's real." What's the clearest problem with that reasoning?
Correct answer: the second option — specificity of the number doesn't equal diagnostic certainty; the finding still needs independent clinical review.
Question 11 · Applied
A patient asks directly whether a computer was involved in flagging a finding on their scan. What is the most appropriate response, per Handout Section F?
Correct answer: the third option — honest disclosure without overstating the tool's authority, per the scripts in Section F.
Question 12
Why does concealing that an AI tool contributed to a finding damage trust, even when the clinician's final judgment was correct?
Correct answer: the first option — the trust damage comes from the concealment being discovered, independent of whether the underlying judgment was right.
Question 13 · Reflection (not scored)
In your own words: describe one AI-assisted finding you might realistically encounter in your own practice, and how you would disclose it to a patient using the framing from Section F.