Ungraded and confidential to you — nothing here is scored or recorded. It just primes you for what's coming. Click an option to see a short note, then move on.
Question 1 · Self-check
Have you already used any kind of AI diagnostic-support, imaging, or ECG-reading tool in your own practice, even informally?
No wrong answer — this just helps you calibrate. If this is new territory, budget extra time for Sections A–C; if you're already using tools like this, you may move faster and focus on Sections D–F.
Question 2
True or false: this module treats an AI-generated differential or imaging flag as the final answer once a clinician has reviewed it.
False — the module's central idea, covered in Handout Section A, is that AI decision support is one voice in the room, never a verdict, however it's reviewed.
Question 3
A tool that scans a chest X-ray and flags a region for closer review, without issuing a final report, is an example of —
The second option — covered in Handout Section C, including a real example directly relevant to this setting: computer-aided detection for TB screening.
Question 4 · Self-check
When you read a colleague's phone consult opinion, do you generally form your own impression first, or do you let their opinion anchor your thinking?
Whatever your habit has been with colleagues, Section D asks you to practice the same discipline with an AI-generated differential — form your own first, on a real worked case.
Question 5
True or false: an AI decision-support tool that performs well overall will automatically rank locally-endemic diseases appropriately, even if its training data came mostly from elsewhere.
False — a tool's ranking reflects the disease prevalence of its training population, not necessarily yours. This "validation gap" is the subject of Handout Section E.
Question 6 · Self-rating
Right now, how confident are you that you'd notice when an AI tool's output doesn't fit your local patient population?
Not at allVery confident
Question 7
When an AI tool flags nothing abnormal on an image, what's the safest way to read that silence?
The second option — the "silent false negative" failure mode covered in Handout Section C.
Question 8
Which of these is the better way to disclose an AI-flagged finding to a patient?
The second option. Handout Section F gives you full scripts for disclosure, disagreement, and patient concerns about AI — you'll use one of your own versions in the live roleplay session.
You're oriented — start the handout
This wasn't graded, and nothing you answered was recorded anywhere. Section D's case exercise is the practical core of this module — don't read the AI-generated differential before writing your own.