# 🤖 Using AI to Organize and Critique, Not Invent

By Better Wise

You have built the raw material: values, strengths map, positioning line, and at least one story told without exaggeration. Now you get a thinking partner that never gets tired and never gets embarrassed to ask a dumb question. The catch is that this partner will happily write fiction about you if you let it. This lesson is about keeping it on the critique side of the line.

By the end you will be able to:

- Explain the difference between using AI as an organizer, a critic, and a ghostwriter
- Write prompts that force AI to work from your material instead of its imagination
- Recognize the four most common ways AI fabricates achievements
- Run a critique pass that produces questions rather than replacement copy

## Three jobs, only two of them yours to keep

Think of AI in three roles, in ascending order of risk.

**Organizer.** You dump your messy notes into it and ask for structure: cluster these values, group these strengths, sort these story beats into a beginning, middle, and end. The content stays yours; AI just moves furniture.

**Critic.** You show it your positioning line and ask what a skeptical hiring manager would challenge. It gives you questions, gaps, and weak spots. You decide what to fix.

**Ghostwriter.** You ask it to write your brand story. This is where portfolios die. What comes back is fluent, generic, and about a fictional person with your name.

Only the first two belong in your portfolio work. **The output of a good AI session is better questions and clearer structure, not better sentences about you.**

> **Note:** 
> Organizer and critic work share one rule: AI can only work with material you provide. If you paste nothing, it invents. If you paste your own evidence, it has something real to push against.

Here is the flow that keeps you safe:

```mermaid
flowchart TD
    A["📝<br/>Your raw material"] --> B["🗂️<br/>AI organizes"]
    A --> C["🔍<br/>AI critiques"]
    B --> D["🧠<br/>You edit and decide"]
    C --> D
    D --> E["✅<br/>Portfolio piece"]
    F["🚫<br/>AI writes from scratch"] --> G["❌<br/>Fiction with your name"]
```

### Exercise 1: Sort the request

You have a paragraph of notes about a volunteer project. Which of these is safe?

1. "Write a compelling paragraph about my volunteer project."
2. "Here is my paragraph. List every claim in it that has no specific evidence."
3. "Turn these five bullet points into a three-beat story, keeping my wording where possible."

**Guidance:** 2 and 3 keep you as the author and AI as the auditor or the sorter. Request 1 hands over authorship, and the model will fill the gaps with plausible-sounding achievements you never had.

## Writing prompts that anchor AI to your evidence

The difference between a useful prompt and a dangerous one is usually one ingredient: a pasted source. Build prompts in four parts.

| Part | What it does | Example |
| --- | --- | --- |
| Source | Pins AI to your facts | "Here are my notes:" plus pasted text |
| Role | Sets the job | "Act as a skeptical reviewer, not a writer" |
| Constraint | Blocks invention | "Use only facts present in my notes" |
| Output shape | Forces questions over copy | "Give me 5 questions I cannot answer yet" |

A full example, ready to adapt:

```text
Act as a skeptical reviewer of personal brand statements.
Use ONLY the facts in the notes below. Do not add achievements,
metrics, employers, or skills that are not written there. If
something is missing, name it as a gap instead of filling it.

Notes:
- Ran a 6-week study group for 8 classmates in a statistics course
- Two people told me the weekly summaries were the reason they passed
- I built the summary template in a shared doc

My positioning line: "I make hard quantitative material usable
for people who think they are bad at math."

Task: List 3 claims in that line my notes do not support,
and ask me one question per gap.
```

Notice what that prompt refuses to do: it does not ask for a better line. It asks for holes.

### Exercise 2: Repair a leaky prompt

"Make my strengths map sound more impressive."

**Guidance:** This prompt has no source, no constraint, and asks for polish. Rewrite it: "Here is my strengths map with the evidence for each item. Point out which strengths have thin evidence and which two claims a reader would question first. Do not reword anything; just list the weak spots." Same material, no invention, more useful output.

## The four ways AI fabricates you

Knowing the failure modes is how you catch them in one read.

1. **Invented metrics.** "Increased engagement by 40%." You never measured it. Numbers appear because they sound credible.
2. **Upgraded titles.** "Led a team" from "coordinated three people informally." Scope inflates quietly.
3. **Borrowed vocabulary.** It reaches for industry-standard phrases ("data-driven," "stakeholder alignment") that sound like everyone else, which is the opposite of positioning.
4. **Smooth causal arcs.** It connects events into a tidy narrative where your choices caused outcomes, when in reality timing and luck were involved. Clean stories are attractive and dishonest.

The check is mechanical: read the AI output line by line and ask of each sentence, "Can I show this?" If the answer is a story about a time it almost happened, delete the sentence.

> **Warning:** 
> Never let a fabricated line survive because it sounds right. Once a claim is in your portfolio, you own it in every interview, and you will be asked to tell that story out loud.

## Running a critique pass

A critique pass is a fixed sequence, not a chat. Do it in this order.

1. Paste your draft and paste the evidence behind each claim as a numbered list.
2. Ask for the three weakest claims and why each is weak.
3. Ask one question per weakness, and answer only the ones you can answer with real evidence.
4. Ask what a reader in your target audience would misunderstand, not what they would dislike.
5. Rewrite the draft yourself, then ask AI to flag any sentence that is not backed by the numbered evidence.

Step 5 is the honesty audit. Expect it to catch one or two rhetorical flourishes, and expect yourself to want to keep them anyway. Delete them.

Good critique prompts sound like this:

- "Which sentence here would a skeptical reader call vague, and what specific detail would fix it?"
- "What am I claiming about my audience that I have not actually verified?"
- "Where does my story skip a step a listener would need?"
- "What is the strongest objection to my positioning line, stated in one sentence?"

### Exercise 3: Your first audit

Take your positioning line from the previous module and run step 2 and step 5 on it with your strengths map as the evidence.

**Guidance:** A healthy result is one or two gaps, not a perfect score. If AI returns no weaknesses, your prompt was too polite, or you did not paste evidence for it to compare against. If it returns a rewritten line, you asked it to write, not to critique. Re-run with the constraint "do not reword my line."

## Field notes from doing this badly first

- **Paste before you ask.** A prompt without your source material is a request for fiction.
- **Ban the word "improve."** Ask for weaknesses, gaps, and questions instead.
- **Keep an evidence column.** Every claim in your portfolio should sit next to the experience that proves it.
- **Watch for tone drift.** After three rounds, AI output starts to sound like AI. If you cannot hear your own voice, start over from your notes.
- **Trust the discomfort.** The question you least want to answer from a critique pass is usually the one worth building your portfolio around.
- **Never paste private data.** No client names, no personal details about other people, no identifying information in your notes.

## What to carry forward

Use AI as an organizer when your notes are messy, and as a critic when your draft feels finished. Keep authorship of every sentence and every claim, and treat any output you cannot evidence as deleted. That discipline is exactly what the next lesson asks of you: you will assemble the full portfolio (values statement, strengths map, positioning line, and one story) and then put it under review, using the critique patterns you just practiced.

Which of these is safe to ask AI to do with your draft?
- a) Rewrite it to sound more professional
- b) List the claims that lack evidence
- c) Invent a stronger opening line

_____

Complete the prompt constraint: "Use only facts present in my _____, and name gaps instead of _____."

A number in AI output that you never measured is an example of an _____.

The final step of a critique pass is an _____, checking each sentence against your numbered evidence.

Front: What are the three roles AI can play in portfolio work?
Back: Organizer, critic, and ghostwriter, with ghostwriter being the risky one.

Front: Which two AI roles belong in authentic portfolio work?
Back: Organizer (structure my material) and critic (question my material).

Front: What single ingredient separates a safe prompt from a dangerous one?
Back: A pasted source: your own notes and evidence.

Front: What are the four parts of an anchored prompt?
Back: Source, role, constraint, and output shape.

Front: What should an anchored prompt ask for instead of better copy?
Back: Gaps, weak claims, and questions you cannot yet answer.

Front: Name the four common ways AI fabricates achievements.
Back: Invented metrics, upgraded titles, borrowed vocabulary, and smooth causal arcs.

Front: What one question tests each AI-generated sentence?
Back: "Can I show this?" If not, delete the sentence.

Front: Why is an invented metric especially dangerous?
Back: You own it in every interview and will be asked to tell that story aloud.

Front: What does the honesty audit step do?
Back: Flags any sentence in your rewrite that is not backed by your numbered evidence.

Front: What result suggests your critique prompt was too polite?
Back: AI returns no weaknesses at all.

Front: What does it mean if AI returns a rewritten positioning line?
Back: You asked it to write rather than critique; re-run with "do not reword my line."

Front: Which word should you ban from prompts to avoid ghostwriting?
Back: "Improve", along with "make it sound impressive" or "more professional".

Front: Why is borrowed vocabulary a problem beyond accuracy?
Back: It makes you sound like everyone else, which defeats positioning.

Front: What data must never go into an AI prompt?
Back: Private or identifying details, including client names and other people's personal information.

Front: What is a normal healthy result of a critique pass?
Back: One or two real gaps, not a perfect score.

---

## Contents

Part of [🧭 Personal Identity and Personal Branding in the Age of AI](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai.md)

- **🪞 Identity Foundations: Values, Strengths, Lived Experience**
  - [🎭 Personal Brand vs Performative Self-Promotion](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/personal-brand-vs-performative-self-promotion.md)
  - [💠 Values Clarification: Naming Your Core Values](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/values-clarification-core-values.md) (sign-in required)
  - [🧩 Strengths Mapping from Lived Experience](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/strengths-mapping-lived-experience.md)
  - [✅ Checkpoint: Identity Foundations](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/checkpoint-identity-foundations.md)
- **🎯 Positioning and Authentic Storytelling**
  - [👥 Audience Analysis: Who You Serve and Why](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/audience-analysis-who-you-serve.md)
  - [📌 Personal Positioning: Claiming a Credible Niche](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/personal-positioning-credible-niche.md)
  - [📖 Authentic Storytelling: Structure Without Exaggeration](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/authentic-storytelling-structure.md)
  - [✅ Checkpoint: Positioning and Storytelling](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/checkpoint-positioning-storytelling.md)
- **🧱 Capstone: Your Authentic Brand Portfolio**
  - [🤖 Using AI to Organize and Critique, Not Invent](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/ai-organize-and-critique-not-invent.md) — this lesson
  - [🗂️ Build Your Personal Brand Portfolio](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/build-personal-brand-portfolio.md) (premium content)
  - [✅ Checkpoint: Brand Portfolio Review](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/checkpoint-brand-portfolio-review.md) (premium content)
- [Course cheatsheet](https://everwise.study/courses/personal-identity-and-personal-branding-in-the-age-of-ai/cheatsheet.md) (premium content)