AIF-C01 Guidelines for Responsible AI Practice Question
An insurance company uses a machine learning model to adjust premiums. During a review, the model is found to be penalizing customers based on zip codes correlated with racial demographics, leading to potential discrimination. Which combination of actions best addresses this fairness issue while maintaining business value?
⚠ Common exam trap
AWS often tests the misconception that removing a single sensitive feature (like zip code) is sufficient to eliminate bias, ignoring that other correlated features can act as proxies and perpetuate discrimination.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Re-engineer features to avoid proxies for protected attributes and rebalance training data
Re-engineering features to remove proxies for protected attributes (e.g., zip codes correlated with race) directly addresses the root cause of bias without discarding all geographic information. Rebalancing the training data helps mitigate skewed representations that could amplify discriminatory patterns, preserving business value by retaining useful predictive signals while aligning with fairness principles under the AIF-C01 Responsible AI guidelines.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Remove the zip code feature from the model and retrain
Why it's wrong here
Dropping zip code alone leaves proxy variables (address, credit history) encoding the same correlation, so bias persists without measuring outcomes. It is tempting because feature removal is a recognised pre-processing mitigation, and would be correct only after bias detection confirms that feature is the sole driver.
- ✓
Re-engineer features to avoid proxies for protected attributes and rebalance training data
Why this is correct
Zip codes act as proxies for protected racial attributes, so removing or re-engineering those features eliminates the discriminatory signal at its source. Rebalancing training data further reduces bias, preserving model utility while addressing the fairness violation.
- ✗
Continue using the model but add a disclaimer about potential bias
Why it's wrong here
A disclaimer leaves the discriminatory pricing intact, so affected customers are still penalised and regulatory exposure remains. It is tempting because transparency is a genuine fairness principle, and would be correct for disclosing model limitations where no protected-group harm occurs.
- ✗
Replace the model with a simpler linear model
Why it's wrong here
A simpler model may still exhibit bias if the data is biased.
Go deeper
Related to this question
About these practice questions
One of 862 original AIF-C01 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AIF-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AIF-C01 exam.