AI-901 ยท Principles of responsible AI
Transparency
Helping users understand AI: disclosing that AI is in use, explaining how outputs are produced, model limitations, confidence scores and interpretability.
Key points
- Transparent AI helps people understand what the system is and how it behaves.
- Tell users when they interact with AI, explain the main factors behind outputs, document limitations, and show confidence or sources where useful.
- Interpretability is the ability to understand why a model produced a result.
Practice questions
Which responsible AI principle says people should understand how AI outputs are produced and what the system's limits are?
- Transparency
- Fairness
- Inclusiveness
- Privacy and security
Show answer
Answer: A. Transparency
Transparency is about clarity on AI use, behaviour and limitations. Fairness concerns equitable outcomes, inclusiveness accessibility, privacy and security data protection.
Which actions help make an AI solution transparent?
- Disclose to users that they are interacting with AI
- Document and communicate the system's known limitations
- Present confidence scores or sources where they are meaningful
- Claim that the AI is always correct
- Leave the system's behaviour undocumented
Show answer
Answer: A. Disclose to users that they are interacting with AI, B. Document and communicate the system's known limitations, C. Present confidence scores or sources where they are meaningful
Disclosure, documented limits and meaningful confidence or sources build understanding. Overclaiming accuracy and skipping documentation undermine it.
A high confidence score guarantees that the model's prediction is correct.
Show answer
Answer: False
False. Confidence is an estimate, not a guarantee; a model can be confidently wrong, so scores should be read together with known limitations.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
