All units
2Foundation Module · Unit 2 of 6
What AI Can and Cannot Do in Clinical Medicine
Draws the precise line between what AI can do in clinical medicine and what it cannot, covering pattern recognition against clinical reasoning, the context AI cannot see, and why its confidence is unrelated to its accuracy.
5 recorded lectures 5 knowledge checks
What you will be able to do
On completing this unit.
- Define AI precisely enough for clinical use, distinguishing narrow task performance from general intelligence, and explain why a vague definition produces unsafe expectations at the bedside.
- Separate pattern recognition from clinical reasoning, and explain why a correlation learned from a dataset is not a cause in an individual patient.
- Identify the context AI cannot perceive — social, environmental and resource constraints, and the implicit signals a clinician reads without being told — and describe how the physician compensates for it.
- Explain why a fluent, authoritative answer can be entirely false, and apply verification discipline rather than judging an output by how it reads.
- Account for the ways training data diverges from the patient in front of you — population mismatch, rare conditions, outdated models — and use individualisation as the safeguard.
AI mirrors its data.
Lectures in this unit
- 2AWhat Do We Mean by AI? (Precise Definition for Safe Practice)
- 2BPattern Recognition vs Clinical Reasoning (Two Very Different Processes)
- 2CContext Blindness In Medical AI (What AI Cannot See)
- 2DConfidence Without Accuracy (Why AI Sounds Right Even When It Is Wrong)
- 2EData Dependency and Bias (AI Reflects Its Data)
This unit is part of the Foundation Module
Six units, 30 recorded lectures, about four hours. The Foundation Module is the entry requirement for every advanced module that follows.