AI Engineer: AI System Design and Ethics Flashcards

What is 'algorithmic bias' in AI systems?

Answer: Systematic and unfair discrimination in AI outputs caused by biased training data or model design

Algorithmic bias occurs when an AI system produces systematically unfair outcomes for certain groups, often reflecting historical biases in training data.

Which principle of responsible AI ensures that stakeholders can understand and interrogate how an AI system makes decisions?

Answer: Explainability and transparency

Explainability means AI decisions can be understood by humans, while transparency ensures the system's design, data, and limitations are disclosed.

What is 'differential privacy' in the context of AI and data science?

Answer: A mathematical framework that adds noise to data or computations to protect individual privacy while enabling aggregate analysis

Differential privacy provides mathematical guarantees that an individual's data cannot be inferred from model outputs by injecting carefully calibrated noise.

In AI system design, what is the purpose of a 'human-in-the-loop' mechanism?

Answer: To include human oversight and intervention in AI decision pipelines for critical or uncertain cases

Human-in-the-loop keeps humans involved in reviewing, correcting, or approving AI decisions, especially where errors have high stakes.

What is 'model cards' documentation in responsible AI?

Answer: Structured documents reporting a model's intended uses, performance across subgroups, limitations, and ethical considerations

Model cards, introduced by Google, standardize transparency by documenting a model's purpose, performance benchmarks, fairness evaluations, and known limitations.

Which US federal principle guides that AI systems should behave reliably and safely across a range of conditions?

Answer: Reliability and safety

Reliability and safety is a core responsible AI principle requiring systems to perform as intended under varied conditions without unintended harmful outcomes.