Half the modules are failures
The AI block spends as much time on where these systems break as on what they do well, which surprises learners expecting either a product demonstration or a warning about robots. Neither is useful to a clinician deciding whether to sign an automated report.
The capability modules
ECG interpretation, where algorithmic detection of reduced ejection fraction and of atrial fibrillation from a sinus tracing genuinely exceeds human performance. Echocardiographic quantification, where automation reduces the inter-observer variability that has always weakened serial comparison. Imaging segmentation. Risk prediction from routine data.
The failure modules
Learners are given outputs that are confidently wrong and asked to work out why. Performance degrades on populations unlike the training data — which matters enormously for learners practising outside North America and Europe, and is demonstrated rather than asserted. Automated systems fail silently. Models trained on past decisions reproduce past bias, including the under-investigation of women.
The accountability module
Closes the block. What you remain responsible for when you act on a model’s output, how to document the reasoning, and how to disagree with an algorithm defensibly. There is little published guidance, so the module is run as structured discussion rather than as teaching.
Open to
All tracks. No technical background is assumed.
PDF, lifetime access, at CardiologyBooks.com.
Reviews
There are no reviews yet.