Gmkt How I Used Ai

---
name: gmkt-how-i-used-ai
description: Writes an honest note on how you used AI in a marketing project, with a task log, the briefs you used, what you checked and changed, what you did yourself, and a three-line version for a CV or interview. Use for "run gmkt-how-i-used-ai", "how do I say I used AI", "should I hide that I used AI", "AI use statement for my portfolio", "explain my AI use in an interview", "AI disclosure for my project", part of the Claude for Marketing Graduates Pack by Polar Bear.
---

# How I Used AI Note

## When To Use
You used AI on your project and do not know whether to hide it or how to say it. This records what you handed to Claude, what you checked, changed and rejected, and what you did yourself, then gives you three lines you can say out loud.

## When Not To Use
If you want one draft checked before it goes out, run the AI Draft Review instead. If the work was assessed by your university or an employer, their rules on AI come first: read them before writing any note, and keep it within your university's rules.

## Inputs
- The project's outputs and, if you kept them, your chats or prompts
- Your AI Draft Review logs or notes on what you changed
- Any rules on AI that apply (course, employer, platform)
If you have none of this, I start from a list of the project's tasks and ask what you did on each, and mark the output as a first draft.

## Approach
The 4D framework (Delegation, Description, Discernment, Diligence) from Anthropic's AI Fluency for Students course gives the four questions this note answers; the framework is named here and its wording stays with its publisher. For public pieces, the Advertising Standards Authority's view on AI disclosure gives the test: would the audience be misled if they did not know AI helped. The failure mode runs both ways: the graduate who hides AI and is caught on a follow-up question, and the one who says "AI did it all" and gives away the thinking that was theirs.

## Workflow
1. Ask three questions: which tasks you used Claude for; one moment where you rejected or changed what came back; which rules on AI apply to this project.
2. Task log: list every project task and mark each as handed to Claude, done with Claude, or done yourself, with one line on why. Planning the brand choice and the final calls usually sit with you; say so if they did.
3. Briefs: paste or summarise the briefs you used for the main tasks (task, audience, voice, constraints, output). A one-line prompt is recorded as one line; no tidying after the fact.
4. Checks: what you checked, changed or rejected and why, with one concrete example from your log (the line before, the line after, the reason). If you have no example, say so and keep a log next time.
5. Your own part and honesty: sources you read yourself, claim checks run, numbers you refused to let Claude fill.
6. Disclosure: for each public piece, ask whether the audience would be misled without saying AI helped. You decide; for anything that is advertising, check with a qualified adviser.
7. Three lines in your words for a CV or interview: what you used AI for, what you checked, what was yours. Read them back aloud; if a line feels like a stretch, cut it.

## Output Format
```markdown
# How I Used AI Note
## Task log
| Task | Handed over / with Claude / myself | Why |
|---|---|---|
| [task] | [choice] | [one line] |
## Briefs I used: [task]: [brief, pasted or summarised]
## What I checked and changed
| What came back | What I did | Why |
|---|---|---|
| [line or output] | [kept / changed to [new line] / rejected] | [reason] |
## What I did myself
[Sources read, claim checks, decisions that were mine]
## Disclosure
| Public piece | Misled without disclosure? | My decision |
|---|---|---|
| [piece] | [yes / no / unsure, why] | [disclose / not needed, check with a qualified adviser if advertising] |
## Three-line version
1. [What I used AI for] 2. [What I checked] 3. [What was mine]
## Decision
[You decide what goes on your CV and in the case study by [date]; where course or employer rules apply, you check them first.]
```

## Done When
- Every project task appears in the log with a choice and a reason
- At least one real example of a change or rejection, or a plain note that none was kept
- Each public piece has a disclosure decision made by you
- The three lines are in your words and you can say them without notes

## Quality Bar
- Only real uses: nothing padded to look skilled, nothing played down to look unaided.
- The 4D framework is named and linked, never quoted.
- No rule on AI is stated as law; disclosure questions end with "check with a qualified adviser".
- Course work stays within your university's rules.
- The note records only what you actually did; Claude never inflates or hides your use of AI.

## Next
Run gmkt-portfolio-post (Portfolio LinkedIn Post) to share the project with the AI note in one line.

## About the makers

This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).