Knowledge check

Quiz

Complete

Test your knowledge.

Evaluate your understanding of regulatory frameworks, risk management, and ML auditing by completing this knowledge check.

Your team is developing an AI system for medical image analysis. A new regulation requires that all AI systems in healthcare must include clear explanations of model outputs for medical professionals. What governance framework or standard would best support this requirement?

A) ISO/IEC 23053 option.B) API access controlsC) Risk register documentationD) NIST Cybersecurity Framework

What is the primary purpose of the NIST AI Risk Management Framework (AI RMF)?

A) To enforce AI fairness standards.B) To provide a structured approach to managing AI risks across the lifecycle.C) To define encryption protocols for sensitive AI data.D) To establish pricing guidelines for AI services.

Which of the following best illustrates a digital supply chain vulnerability in an ML system?

A) Bias in the training data affecting model outputs.B) A third-party API suddenly changing its data format.C) Lack of documentation for model versioning.D) A server room's physical security breach.

A retail company deploys a pricing recommendation model, but after an update, customer complaints surged in one region. An investigation reveals that the model was trained with outdated, incomplete regional data. Which of the following would have best mitigated this risk?

A) A model card explaining the model's purpose and limitations.B) Routine data quality checks and a pre-deployment review of training datasets.C) Encryption of model artefacts at rest.D) Rate limiting access to the pricing API.

A model owner is responsible for ensuring which of the following?

A) The model's logic, fairness assessments, and updates are documented and auditable.B) Third-party API updates are monitored and managed.C) Data engineers define model access permissions.D) Compliance officers test model accuracy.

A logistics company has an ML model that predicts delivery delays. A recent audit finds no documentation of how the model was trained, what features it used, or who approved updates. Which artefact would have helped the team avoid this compliance gap?

A) Risk registerB) Model card and version historyC) Bias audit reportD) API access log

Which of the following is a recommended practice for maintaining audit readiness in an ML system?

A) Waiting to compile documentation until an audit is announced.B) Conducting quarterly compliance reviews and updating documentation regularly.C) Relying on memory to track model changes.D) Only documenting final model versions, not experiments.

You're part of a compliance team reviewing a sentiment analysis model used in customer support. You notice the model struggles with non-English inputs, leading to inconsistent results across languages. What escalation action should the team take?

A) Document the issue and flag it for review in the next scheduled quarterly audit.B) Trigger an immediate escalation, update risk registers, and involve the ML team for a fairness review.C) Reduce the number of languages the model supports to avoid the issue.D) Create a user guide explaining the model's limitations.

Why is embedding compliance practices into daily workflows important for ML teams?

A) It reduces the need for external audits.B) It helps teams quickly resolve issues and demonstrate accountability during an audit.C) It eliminates the need for risk assessment altogether.D) It increases model accuracy across all use cases.

Which artefact is essential for demonstrating model transparency during an audit?

A) Model cardB) Access logC) Risk registerD) API usage reports

Action item: Knowledge check

Complete the quiz below to verify your mastery of this unit's concepts.

Quiz interaction placeholder...

In progress