Responsible AI, Ethics & Career Readiness — Quiz | Data Science using Python

Question 1

What are the four functions of the NIST AI Risk Management Framework?

A Plan, Do, Check, Act

B Govern, Map, Measure, Manage

C Collect, Clean, Model, Deploy

D Train, Test, Tune, Ship

Question 2

Why can a model trained on historical data be biased?

A Models invent bias randomly

B It learns and reproduces discrimination present in the history

C Bias only comes from code bugs

D Data is always fair

Question 3

A disparate-impact ratio of 0.73 suggests…

A perfect fairness

B a possible fairness concern (below the ~0.8 rule of thumb)

C a data error

D high accuracy

Question 4

What is true about the various fairness definitions?

A They are all equivalent

B They often conflict and generally can't all be satisfied at once

C Only one exists

D They are illegal

Question 5

What does SHAP provide?

A Faster training

B An explanation of how each feature pushed a particular prediction

C Data cleaning

D A fairness metric

Question 6

In high-stakes decisions, why might you choose an interpretable model?

A It is always more accurate

B Being able to explain and defend it can outweigh a small accuracy gain

C It needs no data

D It avoids testing

Question 7

Which is a core data-privacy practice?

A Collect as much data as possible

B Minimise collection and pseudonymise/anonymise PII

C Store passwords in code

D Share raw data freely

Question 8

Why is simply removing names insufficient for anonymisation?

A Names are not PII

B Combinations of quasi-identifiers (zip + birthdate + gender) can re-identify people

C It is always sufficient

D Names are encrypted automatically

Question 9

What makes a strong data-science portfolio?

A Many tutorial reruns

B A few end-to-end projects on real problems, clearly documented

C Code with no explanation

D Only Kaggle ranks

Question 10

What is typically tested in a data-science case-study interview?

A Typing speed

B Framing a problem, choosing metrics, outlining data/model and discussing trade-offs

C Memorising library APIs

D Only coding

Question 11

Which role focuses most on production models, MLOps and scale?

A Data Analyst

B ML Engineer

C Research Scientist

D Business Analyst

Question 12

What is a sensible next step after this course?

A Stop learning

B Finish the capstone, publish it, compete/contribute, and specialise

C Only read theory

D Avoid GitHub