AWS Certified AI PractitionerResponsible AIMedium
A credit scoring AI model is found to consistently give lower scores to individuals residing in certain zip codes, even when other financial indicators are similar to those in higher-scoring areas. This pattern is not explicitly programmed but emerges from the training data. What Responsible AI concern does this scenario primarily highlight?
- AAI Robustness
- BAlgorithmic Bias
- CAI Interpretability
- DAI Privacy
Show answer & explanationAnswer & explanation
Correct answer: B. Algorithmic Bias
The scenario describes a systematic and unfair disadvantage to a specific group (residents of certain zip codes) that emerges from the AI model's behavior, even without explicit programming. This is the definition of algorithmic bias, where the algorithm inadvertently perpetuates or amplifies societal biases present in the data.
Why the other options are wrong
- A. AI Robustness deals with the model's stability against attacks or unexpected inputs, not inherent unfairness from data patterns.
- C. AI Interpretability is about understanding how the model makes decisions, but the core issue here is the unfair outcome itself, not just the lack of understanding.
- D. AI Privacy is about protecting personal information, which is not the primary issue described in the credit scoring outcome.
Algorithmic Bias
Algorithmic bias refers to systematic and repeatable errors in a computer system that create unfair outcomes, such as favoring certain groups over others.
- Often arises from biased training data or flawed algorithm design.
- Can lead to discriminatory decisions in various applications.
- Requires careful data auditing, model testing, and mitigation strategies.
Memory trick: Algorithmic bias is when the AI's internal logic, often from its training, builds in unfairness.