AAISM AI Risk Management — Timed 20 Questions

Question 1

An organization is using an LLM-based agent. What is the primary risk associated with 'indirect prompt injection' in this environment?

Accepted Answer

External data retrieved by the agent contains instructions that override system prompts.. Indirect prompt injection occurs when malicious data is embedded in a retrieved document that instructs the LLM to perform unintended actions.

Answer

The model produces hallucinatory facts.

Answer

The API key for the LLM is exposed in the code.

Answer

The training data includes PII of users.

Question 2

Which TWO of the following are primary risks associated with 'membership inference attacks' against an AI model?

Accepted Answer

Exposure of sensitive information about training data subjects.. Membership inference attacks aim to determine if a specific data point was used in the training set, risking privacy leakage.

Answer

Unauthorized modification of model weights.

Answer

Unauthorized access to the model training algorithm.

Answer

Complete system shutdown via resource exhaustion.

Question 3

When utilizing a third-party AI vendor, which contractual requirement is most critical for addressing 'shadow AI' risks within an enterprise?

Accepted Answer

A clause requiring the vendor to disclose all sub-processors.. Standardized procurement processes and inventory tracking (shadow AI) require clear vendor auditing and accountability clauses.

Answer

A mandate to use the vendor's cloud region exclusively.

Answer

A request for a free trial period.

Answer

A requirement for 99.99% uptime SLA.

Question 4

When establishing a risk management framework for GenAI, which THREE factors should be prioritized to satisfy the NIST AI Risk Management Framework requirements?

Accepted Answer

Continuous monitoring of model performance metrics.. NIST AI RMF prioritizes trustworthiness, safety, and reliability through continuous monitoring and governance.

Answer

Requiring all models to be open-source.

Answer

Implementing a strict 'no external API' policy for all models.

Question 5

A machine learning model is showing signs of 'data drift'. What is the most effective initial step in the risk assessment process?

Accepted Answer

Analyze the statistical distribution of input production data versus training data.. Data drift monitoring involves comparing current production data distributions against training data distributions to detect performance decay.

Answer

Retrain the model immediately with all available data.

Answer

Shut down the application to prevent incorrect predictions.

Answer

Change the model architecture to a more complex design.

Question 6

An organization is performing threat modeling for a RAG (Retrieval-Augmented Generation) pipeline. Which specific vulnerability is unique to the retrieval component?

Accepted Answer

Database poisoning within the retrieved document index.. In RAG, if the search or retrieval index is manipulated, the LLM will generate answers based on attacker-supplied context.

Answer

Overfitting of the LLM on the training set.

Answer

Gradient-based evasion attacks on the LLM.

Answer

Prompt injection targeting the LLM model weights.

Question 7

During an adversarial threat modeling exercise for a facial recognition system, you identify a risk of pixel perturbation attacks. Which countermeasure is the most appropriate technical control to implement?

Accepted Answer

Conduct adversarial training by injecting perturbed samples into the training set.. Adversarial training involves training the model on examples containing perturbations, making it robust against those specific attacks.

Answer

Apply data masking to the input images.

Answer

Implement differential privacy during data preprocessing.

Answer

Increase the model epoch count to improve generalization.

Question 8

You are assessing a supply chain risk where an AI model uses a pre-trained base model from a third-party hub. Which action most effectively mitigates the risk of model poisoning during the integration phase?

Accepted Answer

Validate the model hash against the vendor's provided checksum manifest.. Verifying the cryptographic hash of the model weights against the provider's manifest ensures integrity before deployment.

Answer

Implement role-based access control for the training pipeline.

Answer

Run a sensitivity analysis on the model output.

Answer

Enable MFA for the model registry access.

Question 9

A developer is implementing 'Differential Privacy' on a training dataset. What is the primary trade-off they are managing?

Accepted Answer

Privacy budget (epsilon) vs. model utility/accuracy.. Differential privacy adds noise to training data to preserve privacy, which inherently reduces the utility or accuracy of the model.

Answer

Model training speed vs. model cost.

Answer

Security of the training pipeline vs. developer productivity.

Answer

Deployment latency vs. model scalability.

Question 10

What is the primary objective of a 'Bias Audit' in an AI risk management program?

Accepted Answer

To identify disparate impacts on protected populations.. Bias audits ensure that AI models do not produce discriminatory outcomes based on protected classes.

Answer

To verify the model meets legal data storage requirements.

Answer

To reduce the cost of model deployment.

Answer

To ensure the model is computationally efficient.

Question 11

Which of the following best describes 'Model Inversion' as a security risk?

Accepted Answer

An attacker probes the model to reconstruct training samples.. Model inversion involves exploiting model outputs to reconstruct the training data, potentially revealing sensitive information.

Answer

The model is overwritten with malicious logic.

Answer

The model becomes unavailable to users.

Answer

The model predicts the wrong output due to noisy inputs.

Question 12

Which THREE technical controls are recommended to secure an AI model's API endpoint?

Accepted Answer

Enforcing strict schema validation on all inputs.. Securing AI APIs requires defense-in-depth including rate limiting, authentication, and input sanitization.

Answer

Encrypting the model weights at rest in the cloud.

Answer

Removing all logging to prevent log injection.

Question 13

When conducting a risk assessment for an AI system, which TWO of the following are considered 'AI System' components that require individual threat assessment?

Accepted Answer

The training and inference data pipelines.. NIST and ISO standards suggest assessing both the data pipeline and the model environment (inference and training).

Answer

The physical office building security.

Answer

The personal devices of the developers.

Answer

The internal organizational culture.

Question 14

You are implementing 'AI Red Teaming'. What is the most effective way to test against prompt injection?

Accepted Answer

Execute a series of iterative adversarial prompt attempts against the production model.. AI Red Teaming involves systematically probing the model with adversarial prompts to evaluate its response under stress.

Answer

Check for compliance with data privacy laws.

Answer

Review the training data logs for errors.

Answer

Scan the code repository for hardcoded secrets.

Question 15

You are managing risks for a federated learning deployment. What is the primary security concern regarding the 'model updates' sent from the edge clients to the aggregator?

Accepted Answer

Gradient leakage leading to privacy loss.. In federated learning, local model updates can be analyzed to reconstruct private training data from the clients.

Answer

The aggregator becoming a single point of failure.

Answer

Encryption of the communication channel.

Answer

Slow network speeds at the edge.

Question 16

Which THREE criteria are essential for evaluating the risk of an AI vendor's 'Model Card'?

Accepted Answer

Disclosure of training data sources.. Model Cards provide transparency regarding intended use, limitations, and performance, which are critical for risk assessment.

Answer

The vendor's annual financial report.

Answer

The number of employees at the vendor.

Question 17

Which TWO of the following are considered 'AI Supply Chain' risks?

Accepted Answer

Inclusion of poisoned data in third-party datasets.. AI supply chain risks include dependencies on external data and untrusted pre-trained models.

Answer

Lack of internal developer training.

Answer

Excessive use of cloud storage.

Answer

High turnover of AI research staff.

Question 18

A model is vulnerable to 'Evasion Attacks' using FGSM (Fast Gradient Sign Method). Which architectural change most effectively increases robustness?

Accepted Answer

Applying Defensive Distillation.. FGSM uses the model's gradients to craft perturbations; defensive distillation or adversarial training helps mask these gradients.

Answer

Increasing the learning rate during training.

Answer

Switching to a larger neural network.

Answer

Implementing input denoising.

Question 19

What is the most effective control to prevent 'Model Scraping' from an exposed AI prediction API?

Accepted Answer

Implement tiered access and rate limiting.. Rate limiting and request authentication are the standard controls to prevent automated extraction of model knowledge.

Answer

Require users to sign a Terms of Service agreement.

Answer

Increase the complexity of the model.

Answer

Use an obfuscated model file name.

Question 20

When assessing AI third-party vendor risk, which document is most useful for understanding the vendor's data handling practices?

Accepted Answer

The Data Processing Agreement (DPA).. The Data Processing Agreement (DPA) outlines exactly how data is used, stored, and protected.

Answer

The user interface mockups.

Answer

The company's marketing brochure.

Answer

The vendor's quarterly earnings statement.