Executive Summary

This book presents a framework for using generative AI as an assistant in policy and health analytics.

1.7 Core principle

Generative AI assists; analysts decide.

AI may support inquiry, drafting, synthesis, organization, and exploratory analysis, but responsibility for evidence, reasoning, interpretation, and conclusions remains human.

The book develops a governance-aware approach to AI-assisted analytical work that emphasizes: - evidence traceability,
- disciplined review,
- explicit uncertainty,
- accountable authorship,
- and preservation of human judgment.

Rather than treating generative AI as a source of authority or automated decision-making, the book positions these systems as tools that may assist analytical workflows while remaining subject to human review, methodological scrutiny, and institutional accountability.

List of Figures

  • AI-assisted analysis lifecycle
  • Prompt recipe card for inquiry and synthesis
  • Review pipeline for AI-assisted drafts
  • Evidence-to-claim traceability ladder
  • Governance boundary for AI-assisted analysis
  • Failure modes and mitigation practices
  • Checklist → workflow alignment