AAIA Practice Exam 1 — 20 Questions

Question 1

To verify the provenance of a model in Google Cloud Vertex AI, which service should the auditor inspect to review the lineage graph of the artifacts?

Accepted Answer

Vertex AI ML Metadata. Vertex AI ML Metadata tracks the lineage of artifacts and executions within the pipeline.

Answer

Vertex AI Feature Store

Answer

Cloud Logging

Answer

Vertex AI Model Registry

Question 2

When conducting an audit, which THREE of the following represent potential 'Model Risk' areas that require documentation?

Accepted Answer

Susceptibility to adversarial attacks. Adversarial vulnerability, data drift, and bias are three primary risks requiring documentation in an AI audit report.

Answer

The color of the server racks

Answer

The speed of the local internet connection

Question 3

An auditor is evaluating the 'Model Monitoring' dashboard. Which TWO items should the dashboard display to alert the team of potential issues?

Accepted Answer

Real-time model accuracy metrics. Data drift and prediction accuracy are the two most critical metrics for a model monitor.

Answer

The model developer's vacation schedule

Answer

Public stock prices of the firm

Answer

Office building temperature

Question 4

When reviewing an AI system log, what is the 'inference request' ID used for?

Accepted Answer

Audit trail and traceability. The inference ID links the specific request to its model output, facilitating trace-ability during audits.

Answer

Data encryption

Answer

Hyperparameter storage

Answer

Model compression

Question 5

Which metric is commonly used to audit classification models?

Accepted Answer

F1-score. The F1-score balances precision and recall, providing a single metric for classification performance.

Answer

Logarithmic growth

Answer

Mean Squared Error

Answer

CPU utilization

Question 6

Which TWO of the following are essential when performing an 'AI Risk Assessment'?

Accepted Answer

Assessing the likelihood and severity of identified risks. An assessment must identify both technical risks and potential impacts on stakeholders.

Answer

Listing the names of the team members' families

Answer

Choosing the best coffee brand for the office

Answer

Selecting the color scheme for the model's dashboard

Question 7

Which TWO of the following are typical 'Model Risk Management' (MRM) pillars?

Accepted Answer

Independent model validation. MRM is structured around independent validation and continuous monitoring.

Answer

Designing office holiday parties

Answer

Negotiating salaries for the data scientists

Answer

Public relations management

Question 8

Which 'AI Risk Mitigation' strategy is most effective for reducing 'Data Leakage' in training sets?

Accepted Answer

Implementing strict data partitioning and feature selection protocols. Data leakage occurs when information from the target variable leaks into the training features; strict separation and testing prevent this.

Answer

Using more colorful charts in the report

Answer

Increasing the number of servers

Answer

Changing the office seating arrangement

Question 9

Which TWO of the following are legitimate 'AI Governance' concerns regarding third-party AI service providers (SaaS)?

Accepted Answer

Data residency and compliance with local regulations. Auditors must ensure that third-party AI tools comply with internal data handling standards and that the vendor's own governance is adequate.

Answer

The programming language used by the vendor

Answer

Vendor's model update and retraining frequency

Answer

The physical location of the vendor's corporate office

Question 10

When selecting testing techniques for an AI model, which THREE are considered 'Model-Agnostic'?

Accepted Answer

SHAP (SHapley Additive exPlanations). LIME, SHAP, and Partial Dependence Plots (PDPs) are model-agnostic methods that work with any ML model.

Answer

Neural network weight pruning

Answer

Linear regression coefficients

Question 11

What is the key difference between 'AI Ethics' and 'AI Compliance'?

Accepted Answer

Ethics focuses on societal values, while compliance focuses on legal and regulatory requirements. Ethics concerns what *should* be done based on values, while compliance concerns what *must* be done based on rules.

Answer

There is no difference between the two

Answer

Ethics is only for AI, while compliance is for all departments

Answer

Compliance is only for technical roles, while ethics is for executives

Question 12

You are auditing a model's 'Robustness' against adversarial attacks. Which test is most appropriate?

Accepted Answer

Performing adversarial perturbation testing. Adversarial testing involves introducing perturbations to inputs to see if the model's output changes unexpectedly.

Answer

Checking if the server is plugged in

Answer

Checking the model's training accuracy

Answer

Asking the team to describe their favorite movies

Question 13

You are assessing the risk of 'Data Poisoning' in a retail AI model. Which control should you implement during the data ingestion pipeline to mitigate this risk?

Accepted Answer

Input data validation and sanitization. Data validation and sanitization during the ingestion pipeline ensure that malicious or corrupted data samples are blocked before entering the training set.

Answer

Model version control

Answer

Increasing the GPU compute capacity

Answer

Enabling Multi-Factor Authentication (MFA) on the model registry

Question 14

Which THREE criteria are essential when selecting an AI deployment strategy?

Accepted Answer

Service Level Agreements (SLAs). Strategy must balance risk, speed, and resource availability.

Answer

Color of the UI

Answer

Number of coffee machines

Question 15

An auditor finds that a model's 'input feature importance' has changed significantly after a retrain. What is the most appropriate action?

Accepted Answer

Investigate source data and feature logic. Investigate if the change is due to a shift in data or a feature engineering error.

Answer

Ignore the change

Answer

Delete the model

Answer

Upgrade the hardware

Question 16

Which operational process is required to ensure 'Data Privacy' when using user-generated data for model retraining?

Accepted Answer

Data anonymization. Anonymization or pseudonymization must be performed before training.

Answer

Increasing compute power

Answer

Model quantization

Answer

Removing the model

Question 17

When auditing a model deployment pipeline, which TWO aspects are critical to verify to ensure compliance with AI governance frameworks?

Accepted Answer

Validation of data processing against privacy policies. Governance requires both reproducibility and data privacy/security checks.

Answer

Whether the team uses a specific cloud provider

Answer

The speed of the training job

Answer

Usage of proprietary algorithms only

Question 18

When auditing model deployment, what is the primary purpose of 'Shadow Mode' testing?

Accepted Answer

To compare model performance against production traffic without impacting users. Shadow mode allows running the new model in production alongside the old one to compare outputs without affecting users.

Answer

To encrypt model weights in memory

Answer

To hide model artifacts from unauthorized users

Answer

To automate hyperparameter optimization

Question 19

Which of the following is a common 'drift' symptom an auditor should look for in production models?

Accepted Answer

Decreasing model accuracy. When data distribution changes (Concept Drift), the model's accuracy typically degrades over time.

Answer

Increasing server costs

Answer

Increased team size

Answer

Slow Git repository commits

Question 20

What is a 'Model Repository' in an AI audit context?

Accepted Answer

A versioned registry of models. A central, versioned store (like MLflow or Azure Model Registry) for managing trained models.

Answer

A backup hard drive

Answer

A list of developers

Answer

A documentation folder