AWS Certified AI PractitionerResponsible AIEasy
A financial institution is developing an AI model to predict credit default. The data scientists observe that the model consistently performs worse for applicants from a specific demographic group, leading to higher false positive rates for default prediction within that group. Which Responsible AI concept is primarily being violated?
- AAI Robustness
- BAI Transparency
- CAI Privacy
- DAI Fairness
Show answer & explanationAnswer & explanation
Correct answer: D. AI Fairness
AI Fairness addresses the issue of an AI model producing biased or discriminatory outcomes for certain groups, which is precisely what is happening when the model performs worse for a specific demographic.
Why the other options are wrong
- A. AI Robustness deals with an AI model's ability to maintain performance under varied or adversarial conditions, not group-specific bias.
- B. AI Transparency focuses on understanding how AI models make decisions, not the bias in outcomes.
- C. AI Privacy concerns the protection of personal data, which is not the primary issue described.
AI Fairness
AI Fairness ensures that AI systems produce equitable outcomes and do not discriminate against any individual or group.
- Aims to prevent disparate impacts and treatment.
- Addresses biases originating from data, algorithms, or deployment.
- Crucial for ethical and trustworthy AI applications.
Memory trick: Fairness ensures the scales of justice are balanced for all.