Knowledge check

Quiz

Complete

Evaluate your understanding of this unit by completing the knowledge check

These questions summarise core themes from ethical frameworks, privacy law, responsible data practice and governance.

Why is it important to conduct an AI Impact Assessment (AIIA) before deploying an AI model?

A) To maximise model performance without considering fairness or biasB) To assess potential risks, including bias, privacy issues and legal compliance, before deploying AI models.C) To eliminate the need for ongoing monitoring of AI systems after deploymentD) To ensure AI decisions are irreversible once deployed

What is one of the key benefits of Human-in-the-Loop (HITL) systems in AI decision-making?

A) They completely remove human intervention from AI decision-making, making the process fully automated.B) They allow humans to review and override AI-generated decisions in critical applications to ensure ethical compliance.C) They enable AI models to train themselves with no ethical oversight from developers.D) They are only relevant for AI applications in the financial sector and are not widely used in other industries.

Which of the following is an example of an ethical conflict between business objectives and AI ethics?

A) An AI model that improves efficiency while ensuring fairness in decision-making for all users.B) A company that prioritises collecting large amounts of personal data to improve AI recommendations, despite privacy concerns and lack of user consent.C) A government agency using AI to detect fraud while ensuring transparency in decision-making processes.D) A company using AI to streamline hiring while following anti-discrimination laws and ensuring fairness in candidate selection.

What is one way organisations can mitigate bias in AI datasets?

A) Ignoring demographic differences in training data to avoid introducing biasB) Conducting dataset audits and rebalancing training data distributions to ensure fair representation of all groupsC) Collecting as much data as possible, regardless of data quality or fairness considerationsD) Using only historical data without adjustments or fairness interventions

Why is transparency crucial in AI decision-making?

A) It allows stakeholders to understand how AI models make decisions and ensures accountability.B) It reduces AI model efficiency and makes decision-making slower.C) It eliminates the need for human oversight in AI systems.D) It ensures that AI-generated decisions can never be challenged or revised.

What is one of the primary risks associated with black-box AI models?

A) They prevent bias by eliminating human intervention in AI training.B) They make it difficult for stakeholders to understand and justify AI decisions, leading to a lack of accountability.C) They improve AI model accuracy without any ethical trade-offs.D) They eliminate the need for regulation in AI development.

What is a major risk associated with black-box AI models?

A) They improve AI decision-making without any drawbacks.B) They eliminate bias and ensure fairness automatically.C) They make it difficult to interpret and justify AI decisions, reducing accountability.D) They are the only AI models used in high-risk industries such as healthcare and finance.

How does federated learning enhance AI privacy?

A) It allows AI models to be trained on decentralised data without transferring raw user information to a central server.B) It removes the need for any human intervention in AI training.C) It ensures AI models never need updates or retraining.D) It shares all raw user data with central servers for maximum model accuracy.

Why is Ethics-by-Design important in AI development?

A) It integrates ethical considerations at every stage of AI development rather than addressing them as an afterthought.B) It ensures that AI can function without any human oversight.C) It only applies to AI models used in government projects.D) It eliminates the need for transparency in AI systems.

Which principles are essential for ethical AI deployment?

A) Fairness, accountability and transparencyB) Maximising automation and eliminating all human oversightC) Collecting as much personal data as possible to improve model accuracyD) Using AI decision-making without considering its societal impact

Action item: Knowledge check

Answer each question below. Correct answers and short feedback are included for self-review.

In progress