What's changed: In-scope coverage: added security (Macie/Inspector/Secrets Manager) to Ch4 §2 and governance/compliance/cost (CloudWatch/Config/Artifact/Audit Manager/Trusted Advisor/Well-Architected Tool/Cost Explorer/Budgets) to Ch4 §3
4.1Guidelines for Responsible AI
Understand responsible AI dimensions—fairness, explainability, transparency, privacy/safety, robustness—and addressing bias and hallucination.
AI is powerful but carries risks like bias, errors, and privacy harm. Responsible AI is guidance for using AI soundly while addressing these.
4.1.1Responsible AI dimensions
- Fairness: avoid unfair discrimination from biased training data (e.g., a hiring AI not biased by gender).
- Explainability: make the basis for outputs understandable. Transparency: disclose AI use and its capabilities/limits.
- Privacy & safety: protect personal data and prevent harmful/dangerous output.
- Robustness: operate reliably under unexpected input and address hallucination (plausible but wrong output).
Fairness gets special attention because AI bias arises from bias in the training data itself. If past hiring data skews toward certain attributes, an AI trained on it reproduces that skew—not malice, but faithfully learning the bias it was given. Hence the need for diverse, representative data and continuous evaluation/monitoring. In generative AI, hallucination—confidently producing plausible but wrong content—is a major risk; fluency does not equal correctness.
Continue reading — free sign-up
You're reading the free preview. Sign up free to read this section in full, plus every chapter (including 4+) and all questions.