AI governance frameworks in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Key Principle of AI Governance Frameworks

Which of the following is a fundamental principle commonly emphasized in AI governance frameworks to ensure responsible AI use?

A Ignoring data privacy to improve AI accuracy

B Ensuring transparency and explainability of AI decisions

C Prioritizing AI development over human rights

D Maximizing AI system speed regardless of ethical concerns

💡 Hint

Think about what helps users understand AI decisions.

Transparency and explainability help users and stakeholders understand how AI systems make decisions, which is key for trust and accountability.

❓ Model Choice

intermediate

Choosing a Model for Fairness Evaluation

You want to evaluate if an AI model treats different demographic groups fairly. Which approach aligns best with AI governance frameworks focused on fairness?

A Compare model performance metrics separately for each demographic group

B Use a single accuracy metric on the whole dataset

C Ignore demographic data to avoid bias

D Only test the model on the majority group data

💡 Hint

Fairness means checking if all groups are treated equally.

Evaluating performance separately for each group helps detect biases and ensures fairness across demographics.

❓ Metrics

advanced

Interpreting Bias Metrics in AI Governance

An AI governance team uses disparate impact ratio to assess bias. If the ratio is 0.6 for a protected group, what does this indicate?

A The protected group is favored by the AI model

B The ratio indicates no bias

C The AI model has perfect fairness

D The protected group is disadvantaged compared to the reference group

💡 Hint

Disparate impact less than 0.8 usually signals bias.

A ratio below 0.8 means the protected group receives less favorable outcomes, indicating bias.

🔧 Debug

advanced

Identifying Governance Issue in AI Deployment Code

Consider this pseudocode snippet for deploying an AI model: model = train_model(data) if model.accuracy > 0.9: deploy(model) else: deploy(model) What governance issue does this code illustrate?

A It deploys the model regardless of accuracy, risking poor performance

B It correctly prevents deployment of low accuracy models

C It ensures fairness by checking demographic metrics

D It uses explainability checks before deployment

💡 Hint

Look at the conditions for deployment.

The code deploys the model no matter the accuracy, which can cause harm if the model is bad.

🧠 Conceptual

expert

Balancing Innovation and Regulation in AI Governance

Which statement best captures the challenge AI governance frameworks face when balancing innovation and regulation?

A Strict regulations always speed up AI innovation

B No regulation is needed if AI is innovative

C Governance must protect society without stifling beneficial AI advances

D Innovation should be stopped to ensure full safety

💡 Hint

Think about how rules can both help and slow progress.

Effective governance balances protecting people and allowing AI to improve lives.