Exercise · Root-cause matching

When the AI fails, why?

Lesson 4.2

Lesson 4.2 names four conditions behind AI failures. Skim the legend, then for each scenario decide which condition best explains what went wrong.

The four conditions
Pattern-based learning
Extending a learned pattern into cases where it no longer fits.
Uneven training data
Some groups, languages, or cases are sparse or missing.
Weak grounding
Output isn\u2019t checked against verified reality before being produced.
Drift
The environment changed after the model was trained or deployed.
Five cases
0 / 5 correct