AAISM Sample Questions & Answers
Controls over AI model lifecycle and its data management, plus AI security architecture, take the top weight, tied closely with AI risk assessment, threats and vendor management, and governance frameworks covering strategy, policy and the AI asset and data lifecycle.
Launch the full AAISM simulator →Showing 10 of 20 free samples.
- Question 1Intermediate
AI Risk Management · AI Threat and Vulnerability Management
A security analyst discovers that a deployed image recognition model has been misclassifying images that contain a specific, imperceptible pattern of noise overlay. This pattern appears to have been intentionally crafted. This is an example of which type of adversarial attack?
Show answer & explanation
Correct answer: D
This describes an Evasion Attack (often using adversarial examples). The attacker manipulates the input data (inference time) with perturbations (noise) to cause the model to make a prediction error. It does not tamper with the training data itself.
- Question 2Beginner
AI Technologies and Controls · Privacy, Ethical, Trust and Safety Controls
The ___________ is a document that provides a high-level description of the AI model, including its intended use, limitations, training data provenance, and performance metrics, serving as a key transparency control.
Show answer & explanation
Correct answer: C
A Model Card (introduced by Google researchers and widely adopted) is a standardized document that details performance characteristics, intended use, limitations, and ethical considerations of an AI model.
- Question 3Advanced
AI Risk Management · AI Vendor and Supply Chain Management
When integrating a third-party AI model into a critical business process, which of the following is the MOST effective method to mitigate supply chain risks related to hidden backdoors or malicious code in the model artifacts?
Show answer & explanation
Correct answer: D
Cryptographic signing of model artifacts (e.g., using Sigstore or similar tools) ensures integrity and authenticity. Validating the signature before loading prevents the execution of tampered models or models from untrusted sources, directly addressing the risk of supply chain injection attacks.
- Question 4Beginner
AI Technologies and Controls · Privacy, Ethical, Trust and Safety Controls
True or False: In the context of AI security, 'Explainability' (XAI) is purely an ethical requirement and has no impact on security incident investigation.
Show answer & explanation
Correct answer: B
False. Explainability is critical for security forensics. If an AI system behaves maliciously (e.g., due to an adversarial attack or poisoning), security analysts need explainability tools (like SHAP or LIME) to understand why the model made that decision and identify the root cause of the breach.
- Question 5Intermediate
AI Risk Management · AI Threat and Vulnerability Management
An organization uses a Large Language Model (LLM) to summarize customer support tickets. Security testing reveals that by crafting a specific prompt, an attacker can force the model to ignore its safety guidelines and reveal its initial system instructions. What is this vulnerability called?
Show answer & explanation
Correct answer: D
Prompt Injection (specifically direct injection or jailbreaking) involves manipulating the input to override the model's original instructions or safety constraints. This allows the attacker to control the model's output.
- Question 6Intermediate
AI Governance and Program Management · AI Asset and Data Life Cycle Management
You are implementing a 'Shadow AI' discovery program. Which of the following technical methods would yield the MOST accurate inventory of unauthorized AI services being used by employees?
Show answer & explanation
Correct answer: A
Analyzing network traffic via CASB or Secure Web Gateways (SWG) is the most reliable technical method. It identifies actual traffic to known AI service domains (e.g., openai.com, huggingface.co) regardless of whether employees have reported them.
- Question 7Advanced
AI Technologies and Controls · AI Security Architecture and Design
Case Study:
AlphaCorp is deploying a fraud detection model trained on sensitive customer transaction data. The data science team wants to use an external cloud provider for training due to compute constraints. The CISO is concerned about the cloud provider accessing the raw data in memory during training.
Which technology should the CISO recommend to ensure the data remains encrypted even while it is being processed in the cloud provider's RAM?
Show answer & explanation
Correct answer: A
Confidential Computing uses hardware-based Trusted Execution Environments (TEEs) or enclaves (like Intel SGX, AWS Nitro Enclaves) to isolate data in memory. This ensures that even the cloud provider's hypervisor or administrators cannot view the data while it is being processed (encryption-in-use).
- Question 8Intermediate
AI Technologies and Controls · AI Life Cycle (Model Selection, Training, and Validation)
In an MLOps pipeline, 'Data Drift' is a significant risk that can lead to security degradation or model failure. Which of the following best describes Data Drift?
Show answer & explanation
Correct answer: B
Data Drift (or Covariate Shift) occurs when the distribution of input data in production diverges from the distribution of the training data. This can cause the model's accuracy to degrade and may indicate an adversarial attack or a change in the environment.
- Question 9Intermediate
AI Technologies and Controls · Privacy, Ethical, Trust and Safety Controls
Which of the following metrics is commonly used to measure the 'fairness' of an AI model by checking if the prediction accuracy is consistent across different demographic groups?
Show answer & explanation
Correct answer: C
Equalized Odds is a fairness metric that requires the model to have equal true positive rates and false positive rates across different protected groups (e.g., gender, race). It ensures that the model is equally accurate for all groups.
- Question 10Beginner
AI Risk Management · AI Vendor and Supply Chain Management
A software company is developing a new AI feature. The security team wants to ensure that the dependencies used in the ML pipeline (e.g., Python libraries like NumPy, PyTorch) are free from known vulnerabilities. Which document should they request and analyze as part of the supply chain security process?
Show answer & explanation
Correct answer: D
A Software Bill of Materials (SBOM) is a comprehensive list of all components, libraries, and dependencies used in a piece of software. Analyzing the SBOM allows the team to identify if any included components have known vulnerabilities (CVEs).
Ready for the real thing?
The full AAISM simulator has every exam-style question, timed mode, and instant scoring.