How to use this book

We begin with a short orientation. In this chapter we set out the daily cadence, the chapter template, and the conventions used throughout, so that the reader knows what to expect.

The daily cadence

Each day has the same shape:

  • Lecture content (~1 hour). Read the chapter, trying each prompt against an AI tool of your choice as you encounter it. Do not skip the exercises; the chapter is designed to be read with an AI assistant open and a real (or scratch) research task at hand.
  • Homework (~2 hours). Five to eight problems, with worked solutions at the end of the chapter. Attempt each problem before checking the solution.

The cadence assumes one chapter per day for five consecutive days. The workshop may be compressed (two chapters per day across two-and-a-half days) or expanded (one chapter per week across a month) depending on the reader’s other commitments; we leave the choice to the reader.

Chapter template

Each chapter follows the same template, and we shall keep to it consistently:

  1. Learning objectives. A bulleted list of what the reader should be able to do by the end of the day.
  2. Lecture. The substantive content for the day, with prompts to try and outputs to evaluate. Each major subsection ends with a short Check your understanding callout.
  3. Further reading. The primary sources behind the chapter’s claims: papers, court records, incident reports, and policy documents, so the reader can verify a claim rather than take the book’s word for it.
  4. Worked example. A small but realistic research scenario that uses the day’s content end to end.
  5. Homework. Problems organized from easier to harder.
  6. Solutions. Worked solutions to all homework problems. Read only after attempting the problem.

There are no examinations in this book. The per-subsection check-your-understanding callouts are low-stakes comprehension checks, not graded quizzes; the homework solutions remain the primary vehicle for self-assessment.

A note on scope

This book concerns the effective, efficient, and responsible use of existing AI tools within an active research project, and we mean that in order of emphasis. Each day leads with a concrete technique for getting research work done faster, and follows it with the verification and disclosure judgment that keeps the speed from costing more time than it saved.

We do not teach how such tools are built, and we treat no single vendor’s tool as the subject. Both the efficiency techniques and the verification judgment presented here apply regardless of which large language model or coding assistant the reader adopts. Where an example names a specific tool (Claude, ChatGPT, Claude Code, GitHub Copilot), it does so for concreteness, not endorsement.

Conventions

The visual cues used throughout the book are described on the Conventions page. In brief:

  • A command the reader types follows a $ prompt; output has no prompt.
  • Prompts sent to an AI tool appear in prompt blocks; tool responses appear in response blocks.
  • Claims about AI output are marked verified or unverified.
  • Placeholders the reader replaces are written in <angle-brackets>.

Companion volumes

This workshop sits within the rgtlab curriculum sequence. The two preparatory boot camps precede it, and two related volumes extend its reproducibility material:

Day 5 describes how the provenance and disclosure material here connects to Git and GitHub for Biostatistics and to Reproducible Research for the Health Sciences.