奇创岛

From «How to Read a Book»

Build an AI-Assisted Analytical Reading Toolkit

You'll design a reusable AI prompt toolkit grounded in Adler's three stages of analytical reading (find the skeleton → interpret the content → critique the author) — one or two prompts per stage covering scenarios such as having AI surface core arguments, test your comprehension, and list counterexamples — plus a boundary statement for each.

Final work

AI-Assisted Analytical Reading Toolkit

Estimated time

1–2 hr

Submitted

…

Your final work

Purpose:Translate Adler’s analytical reading method into an on-demand AI prompt system so that AI acts as a ‘question mirror’ and ‘comprehension inspector’ during reading — not as a ghostwriter who thinks for you.

Parts:

  • Four-level reading System Prompt template (with usage-boundary notes)
  • 4 Core-Questions Prompts (What is the book about / How is it developed / Is it true / What does it mean to me)
  • Argument-Skeleton Extraction Prompt
  • Consensus-Check Prompt (reach agreement with the author; distinguish words from terms)
  • Syntopical-Reading Cross-Book Integration Prompt
  • Genre-Specific Rules Prompt (at least 2 genres)
  • AI Usage-Boundary Declaration (what AI can do vs. what it must not replace)
  • Toolkit User Guide (when to use each prompt and how to operate it)

Use cases:

  • · Full AI-assisted reading of your next deep-read book
  • · Cross-book synthesis of views during syntopical reading
  • · Discussion-preparation tool for co-reading with others
  • · Personal self-check instrument for gauging reading depth

Pick a topic

Pick the topic closest to you, or write a custom one when you submit.

Personal Life

Learning & Growth

Work & Projects

Communication & Relationships

Tools you'll use from the book

Four-Level Reading System Prompt

Sets up the AI in the role of ‘analytical reading assistant,’ telling it which level you’re working at — foundational, inspectional, analytical, or syntopical — and constraining it to act only within that level.

How to use it here:

Paste this System Prompt before each AI interaction to establish your current reading level, so the AI supports you at that level rather than jumping ahead to summarize or judge on your behalf.

Boundaries:

AI must not complete your comprehension task for that level. The System Prompt should include constraints such as ‘do not draw conclusions for me — only help me ask questions and check my understanding.’

4 Core-Questions Prompts

Corresponds to Adler’s four core active-reading questions: What is the book about as a whole? How is it developed? Is it true? What does it mean to me? — one standalone prompt per question.

How to use it here:

After finishing a chapter or book, use the four prompts in sequence to have AI check whether your own answers are complete — not to have AI answer the questions for you.

Boundaries:

Write your own answer first, then feed it to AI for review. You must not skip writing your own answer and ask AI directly.

Argument-Skeleton Extraction Prompt

Asks AI to extract a three-level skeleton — main proposition → supporting propositions → key evidence — from the chapter content you provide, and to flag logical leaps or insufficient evidence.

How to use it here:

Use during the analytical-reading stage (Adler Rules 11–13) to evaluate the author’s argument. First outline the skeleton yourself, then use AI to cross-check for anything you missed.

Boundaries:

The skeleton AI extracts is a reference, not a final answer. The ultimate judgment of whether the author’s argument holds must be made by you.

Consensus-Check Prompt

Asks AI to verify whether your understanding of a term or proposition matches the author’s intent, distinguishing between its everyday meaning and the author’s specific usage.

How to use it here:

When reading philosophy, social science, or business books, write out your understanding of a key term (e.g. ‘reading,’ ‘understanding,’ ‘argument’) first, then use this prompt to have AI cross-check it against the original text.

Boundaries:

You must not ask AI to ‘just explain what this word means.’ You must provide your own interpretation for AI to compare against — otherwise the check loses all value.

Syntopical-Reading Prompt

Used for cross-book synthesis: asks AI to display each book’s answer to a shared question side by side, noting similarities and differences, without acting as the judge.

How to use it here:

During the ‘defining the issue’ and ‘analyzing the discussion’ steps of syntopical reading, paste relevant passages from each book and ask AI to produce a ‘comparison table of each book’s position’ — not a synthesized conclusion.

Boundaries:

The comparison table should only present each book’s actual position. AI must not add a ‘best answer’ or ‘synthesized conclusion’ to the table. The judging is your job.

Work rules

Your work MUST include

  • Must include at least one immediately usable prompt written for a real book you are about to read or are currently reading
  • Every prompt must contain a concrete input template and a specified output format
  • Must state at least 3 AI usage boundaries (what AI can do and what it should not do)
  • Must explain the overall sequence and appropriate timing for using the toolkit
  • Must cover at least two of the three analytical-reading stages (find the skeleton / interpret the content / critique the author)

Your work CANNOT just be

  • Must not be a reading-response essay that merely summarizes Adler’s methodology
  • Must not list only ‘have AI summarize this for me’-style prompts that outsource thinking to AI
  • Must not omit usage-boundary statements
  • Must not lack examples tied to a specific book or scenario
  • Must not turn the AI toolkit into a plan where ‘AI reads the book so I don’t have to’

AI can help you here

Round 1: Help me choose the reading scenario that fits me best

When to use: You’re not yet sure which book you want to use this toolkit with, or you’re unsure which reading level you should be working at.

I am working on the ‘{{route name}}’ project using *{{book title}}*.

Based on my situation, please help me decide which reading level I should choose (inspectional / analytical / syntopical) and the 1–2 most suitable prompt scenarios for where I am right now.

My situation:
[Fill in: which book you want to read, how far along you are, what problem you want to solve, how much time you have]

Please output:
1. Recommended reading level and the reason
2. The 1–2 prompt scenarios that best fit me
3. What I need to prepare before entering this scenario
4. Which steps should not involve AI (kept for my own thinking)

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Round 2: Help me design a specific AI prompt

When to use: You’ve already chosen your book and reading level, and now you want to turn a specific analytical-reading step into a usable prompt.

I am working on the ‘{{route name}}’ project using *{{book title}}*.

My topic is:
{{topic}}

Please help me turn the following Adler analytical-reading step into a concrete, usable AI prompt. Requirements:
1. Include a System Prompt setting (the AI’s role and constraints)
2. Include a user-input template (what information I should fill in)
3. Include a desired output format (what structure AI should produce)
4. Include a usage-boundary statement (what AI must not do)

The step I want to convert:
[Fill in: e.g. ‘Adler Rule 11: identify the author’s core propositions’, or ‘Core Question 3: Is this book true?’]

Please output:
- Full prompt text
- Usage instructions (step-by-step)
- Boundaries (2–3 items)
- What I must complete on my own before using this prompt

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Round 3: Help me check whether the complete toolkit is ready to use

When to use: You’ve finished your draft toolkit and run at least one test, and you want to verify its quality before submitting.

I am about to submit my Shufang Island project work.

Book: *{{book title}}*
Project route: {{route name}}
My topic: {{topic}}

My draft work (AI-Assisted Analytical Reading Toolkit):
{{draft work}}

Please check this toolkit against the following criteria:
1. Does each prompt genuinely correspond to a specific stage or rule in Adler’s analytical reading?
2. Are there any prompts that, in practice, would cause ‘AI thinks for me’ rather than ‘AI tests my thinking’?
3. Are the usage-boundary statements clear, and do they cover the most easily misused scenarios?
4. Does the toolkit as a whole form a coherent reading-support system (rather than a pile of unconnected prompts)?
5. Does the test-run result demonstrate the toolkit’s real-world effectiveness?

Please output:
- Overall assessment
- Quality evaluation of each prompt (one sentence each)
- Changes that must be made
- Areas that could be strengthened
- A revised structural suggestion for the complete toolkit

Yellow placeholders need you to fill in before using the AI.

AI can help you organize ideas, but cannot make final judgments for you. Don't let AI fabricate experiences, cases, or misleading content.

Submit my work