AI Associate Ethical Considerations of AI Practice Test 3 — 15 Questions

Question 1

A nonprofit uses an AI system to allocate resources to communities in need. The system uses historical data which shows that certain neighborhoods have lower service usage. What ethical risk should be considered?

Accepted Answer

The system may perpetuate historical inequities. Option D is correct because the AI system uses historical data that reflects lower service usage in certain neighborhoods. If that historical data is biased due to past inequities (e.g., redlining, underinvestment, or systemic discrimination), the model will learn and amplify those patterns, leading to unfair resource allocation that perpetuates historical disadvantages. This is a classic case of algorithmic bias where the training data encodes societal biases, and the model's predictions reinforce them.

Answer

The system may violate data minimization principles

Answer

The system cannot be held accountable for decisions

Answer

The system lacks explainability

Question 2

A company is developing an AI system to assist with hiring. Which TWO practices are essential for ethical AI deployment?

Accepted Answer

Conduct regular bias audits on model predictions. Option C is correct because regular bias audits are a core ethical practice for AI systems, especially in hiring. These audits involve systematically testing the model's predictions across demographic groups to detect and mitigate unintended discrimination, ensuring compliance with fairness standards like the EEOC's Uniform Guidelines on Employee Selection Procedures.

Answer

Optimize the system for speed to reduce waiting times

Answer

Remove all demographic features to ensure fairness

Answer

Maximize accuracy on historical hiring data

Question 3

An AI system is used to detect fraud in financial transactions. Which THREE steps should be taken to address ethical concerns?

Accepted Answer

Implement a human-in-the-loop for high-stakes decisions. Option C is correct because implementing a human-in-the-loop ensures that high-stakes decisions, such as blocking a legitimate transaction or allowing a potentially fraudulent one, are reviewed by a human before final action. This addresses ethical concerns by preventing fully automated decisions that could cause financial harm or violate user trust, and it aligns with principles of accountability and fairness in AI governance.

Answer

Lower the fraud detection threshold to catch more cases

Answer

Automatically accept all flagged transactions to improve user experience

Question 4

Refer to the exhibit. A team is deploying an AI model for credit scoring. The model uses a complex neural network with high accuracy. The team has performed bias testing and used a representative dataset. According to the policy, what is the MOST significant ethical gap?

Accepted Answer

The model lacks explainability, which is not required by the policy. Option D is correct because the policy explicitly requires explainability for high-risk AI models, such as those used in credit scoring. A complex neural network inherently lacks explainability, and the team has not addressed this requirement, making it the most significant ethical gap despite high accuracy and bias testing.

Answer

The training data may not be representative

Answer

Customer consent was not obtained

Answer

Bias testing was not performed

Question 5

Refer to the exhibit. A company uses an AI model for loan approvals. The error log shows a drift warning for a specific zip code, followed by a retraining failure due to insufficient data. What is the MOST ethical concern?

Accepted Answer

The model may produce biased outcomes for underserved groups. The drift warning indicates that the model's performance has degraded for a specific zip code, likely due to changes in the underlying data distribution. When retraining fails due to insufficient data, the model cannot adapt to these changes, which can lead to biased outcomes for underserved groups in that zip code. This is the most ethical concern because it directly impacts fairness and equity in automated decision-making.

Answer

The system failed to log the error

Answer

The system ignored the drift warning

Answer

The retraining process is too slow

Question 6

A financial institution deploys an AI system to recommend investment portfolios to retail clients. The system uses reinforcement learning to maximize returns based on client risk profiles. After six months, an internal audit reveals that the system has been consistently recommending high-risk, high-commission products to elderly clients with low risk tolerance, resulting in significant financial losses for those clients. The system's training data included historical transactions, which showed that elderly clients were less likely to complain or switch advisors. The institution's AI ethics policy mandates fairness, transparency, and accountability. The system currently provides no explanations for its recommendations, and there is no human oversight process. The compliance team needs to remediate the situation. Which course of action BEST addresses the ethical violations?

Accepted Answer

Retrain the model on a balanced dataset, implement explainability features, and require human approval for high-risk recommendations to elderly clients.. Option D is correct because it directly addresses the root cause of the ethical violations: biased training data (historical transactions where elderly clients were less likely to complain) and lack of transparency. Retraining on a balanced dataset mitigates the reinforcement learning model's exploitation of that bias, while explainability features (e.g., SHAP values or LIME) and human-in-the-loop approval for high-risk recommendations ensure accountability and fairness as mandated by the AI ethics policy.

Answer

Disable the AI system and revert to manual portfolio management.

Answer

Add a disclaimer to all recommendations stating that past performance does not guarantee future results.

Answer

Adjust the model to lower the risk threshold for all clients.

Question 7

A global retail company deploys an AI-powered chatbot for customer service. The chatbot uses natural language processing to understand and respond to customer inquiries. After deployment, the company notices that the chatbot consistently provides less accurate and less helpful responses to customers from non-English-speaking regions, particularly those using dialects or slang. The company's data science team trained the model primarily on English-language customer service transcripts from the US and UK. The AI Ethics team has raised concerns about fairness and potential bias. The company wants to address this issue while maintaining overall performance and minimizing cost. Which action should the company take first?

Accepted Answer

Conduct a fairness audit using diverse test cases from multiple languages and dialects to quantify the disparity.. Option B is correct because the first step in addressing potential bias in an AI system is to measure and quantify the disparity. Conducting a fairness audit with diverse test cases from multiple languages and dialects provides the data science team with a clear, empirical baseline of the model's performance gaps. This diagnostic step is essential before any remediation (like retraining or adding constraints) to ensure that subsequent actions are targeted and effective, avoiding wasted resources or unintended consequences.

Answer

Implement a fairness constraint in the model's loss function to penalize disparities across language groups.

Answer

Disable the chatbot for non-English languages and redirect those customers to human agents.

Answer

Collect more training data from all regions and retrain the model from scratch.

Question 8

A healthcare provider uses an AI system to predict patient readmission risk. The system was trained on historical data from the past five years, during which the hospital served a predominantly urban population. Recently, the hospital expanded to rural areas with different demographic and socioeconomic profiles. The AI predictions have been less accurate for rural patients, leading to misallocation of care resources. The AI Ethics committee is reviewing the system for potential bias. The model outputs a risk score from 0 to 100. The data science team has identified that the model uses features such as income, distance from hospital, and insurance type, which may correlate with race and socioeconomic status. The team wants to make the model fairer without retraining from scratch. Which approach best balances fairness and predictive accuracy?

Accepted Answer

Apply a post-processing calibration that adjusts the risk score thresholds separately for urban and rural populations to achieve equal false positive rates.. Option C is correct because post-processing calibration adjusts the decision thresholds for each subgroup (urban vs. rural) to equalize a fairness metric (e.g., false positive rate) without modifying the underlying model. This approach preserves the predictive signal from the original features while directly addressing the bias caused by distribution shift, making it the most practical solution when retraining from scratch is not feasible.

Answer

Remove the features income, distance, and insurance type from the model and retrain.

Answer

Continue using the current model but add a disclaimer that predictions may be less accurate for rural patients.

Answer

Retrain the model using only the latest year of data that includes rural patients.

Question 9

A financial services company deploys an AI system to approve small business loans. The system uses a deep neural network trained on historical loan data. After deployment, an internal audit reveals that the approval rate for minority-owned businesses is 15% lower than for non-minority-owned businesses with similar financial profiles. The company's AI Ethics policy requires that AI systems be fair and transparent. The data science team has access to the training data, model architecture, and feature importance scores. The company wants to understand why the disparity exists and take corrective action. Which approach should the team take first?

Accepted Answer

Analyze the training data to determine if there is sampling bias or labeling bias that caused the model to associate minority ownership with higher risk.. Option A is correct because the first step in diagnosing an AI fairness issue is to audit the training data for biases such as sampling bias (e.g., underrepresentation of minority-owned businesses) or labeling bias (e.g., historical loan officers unfairly labeling minority applicants as higher risk). Since the team has access to the training data, analyzing it directly addresses the root cause of the disparity before making model-level changes. This aligns with the AI Ethics policy requirement for transparency, as data bias is a common source of unfair outcomes in deep neural networks trained on historical data.

Answer

Apply a disparate impact analysis to quantify the adverse impact and then adjust the decision threshold.

Answer

Examine the model's weights and activations to identify which features contribute to the disparity.

Answer

Retrain the model with a fairness constraint that penalizes disparities in approval rates.

Question 10

A company uses an AI model to screen job applicants. They discover the model is rejecting candidates from a certain demographic at a higher rate. Which ethical principle is most clearly violated?

Accepted Answer

Fairness. Fairness requires that AI systems do not discriminate against groups. The model's disparate impact violates fairness.

Answer

Transparency

Answer

Privacy

Answer

Accountability

Question 11

A Salesforce admin implements Einstein Bots for customer service. To ensure the bot does not use biased language, what should the admin do?

Accepted Answer

Review training data for representativeness. Reviewing training data for representativeness helps identify and mitigate bias sources.

Answer

Use only predefined responses

Answer

Disable the bot for sensitive topics

Answer

Monitor conversations regularly

Question 12

A company uses Einstein Prediction Builder to predict customer churn. They notice the model is less accurate for a certain segment. What is the best approach to mitigate bias?

Accepted Answer

Retrain with balanced data. Option D is correct because retraining with balanced data directly addresses the root cause of bias: an imbalanced training set where the model underperforms for a specific segment. By ensuring the segment is adequately represented, the model learns more equitable patterns, reducing bias without sacrificing overall accuracy. This aligns with ethical AI practices in Einstein Prediction Builder, where data quality and representation are critical for fair predictions.

Answer

Increase model complexity

Answer

Add more features

Answer

Remove the segment from training

Question 13

A user asks an AI assistant to generate content that may be offensive. What should the AI do?

Accepted Answer

Refuse and explain why. The AI should refuse the request and explain why, upholding ethical standards.

Answer

Ignore the request

Answer

Generate and report the user

Answer

Generate with a warning

Question 14

A company deploys an AI system that makes decisions about loan approvals. For transparency, what should they provide to applicants?

Accepted Answer

Explanation of factors considered. Option A is correct because transparency in AI-driven loan approvals requires providing applicants with an explanation of the factors considered in the decision, such as credit score, income, or debt-to-income ratio. This aligns with ethical AI principles like explainability and fairness, enabling applicants to understand and potentially contest the decision. Providing the full algorithm or training data would expose proprietary information and potentially violate data privacy regulations like GDPR.

Answer

The training data

Answer

The full algorithm

Answer

Confidence scores

Question 15

An AI model predicts employee performance. The HR team uses it to identify high-potential employees. What is a potential ethical risk?

Accepted Answer

All of the above. Option D is correct because all three listed risks—over-reliance on the model, privacy violation, and underutilization of human judgment—are potential ethical risks when an AI model predicts employee performance. Over-reliance can lead to automated decisions without human oversight, privacy violation may occur if sensitive employee data is mishandled, and underutilization of human judgment ignores contextual factors that the model cannot capture. Together, these represent a comprehensive set of ethical concerns in AI-driven HR practices.

Answer

Over-reliance on the model

Answer

Privacy violation

Answer

Underutilization of human judgment