1 Introduction

Generative artificial intelligence systems are increasingly used to draft, summarize, and explain complex material. They are embedded in everyday tools and often described as assistants, copilots, or collaborators. Despite their rapid adoption, there is little shared understanding of what it means to use generative AI well in research, policy, and analytical contexts.

This book approaches generative AI as an assistant to thinking, not as a source of authority. Rather than treating AI outputs as answers, it treats them as prompts for reflection: ways to surface assumptions, explore alternative framings, clarify arguments, and identify gaps in reasoning or evidence. The central premise is that generative AI can support analytical work without replacing human judgment, responsibility, or methodological care.

In research and policy settings, the risks of uncritical AI use are substantial. Fluent language can obscure uncertainty, fabricated details can appear plausible, and the line between evidence and interpretation can blur. These risks are not failures of individual users alone; they follow directly from how generative models produce text. Recognizing this, the approach taken here emphasizes deliberate use, systematic review, transparency, and epistemic humility.

Throughout this book, generative AI is treated as one tool among many in the analytical process. It can be helpful for organizing ideas, synthesizing material, and improving clarity, but it cannot evaluate evidence, determine truth, assess risk, or assume accountability. Responsibility for interpretation and conclusions always remains with the human analyst or researcher.

This book is written for readers who are curious about generative AI but cautious about its implications: researchers, policy analysts, health system analysts, students, and others engaged in inquiry and investigation. Its goal is not to promote efficiency or automation, but to support thoughtful, disciplined engagement with generative AI as an assistant—used critically, provisionally, and in service of clearer thinking.

1.1 What this book is

This book is a practical and conceptual guide to using generative AI as a research and analysis assistant. It focuses on how large language models can support inquiry and analytical work by helping with:

  • Structuring questions and lines of inquiry
  • Exploring alternative framings or interpretations
  • Summarizing and synthesizing existing material
  • Making assumptions, uncertainty, and trade‑offs explicit
  • Improving clarity in analytical and explanatory writing
  • Reviewing and revising AI‑assisted drafts responsibly

It emphasizes how AI outputs are evaluated, revised, governed, and integrated into analytical work—not just how they are generated.

1.2 What this book is not

This book does not treat generative AI as:

  • A source of factual authority or evidence
  • A decision‑maker, evaluator, or validator
  • A replacement for subject‑matter expertise or peer review
  • An autonomous research or analysis system

Throughout the book, responsibility for accuracy, interpretation, judgment, and impact remains fully with the human user.

1.3 How to use this book

Readers may engage with this book in different ways:

  • Sequentially, as a conceptual guide to AI‑assisted research and analysis
  • Selectively, by focusing on specific patterns, examples, or review practices
  • Reflectively, as a way to examine their own use of generative AI in analytical work