The Creation Log
Why This Matters: James and the Attribution Problem
James had his Draft 1 open in one window and an AI tool in another. His fingers were already on the keyboard, prompt half-typed, when he stopped himself.
"Hang on." He looked at Emma. "If I just paste my entire draft and say 'improve this,' I'll get back a polished version. But I won't know which improvements were mine and which were the AI's."
Emma raised an eyebrow. "So?"
"So in two hours, when someone asks me what I contributed, I'll be standing there like the guy in the group project who 'helped with the PowerPoint.'" He leaned back. "At my old company, we had this problem with consultants. They'd come in, rewrite our processes, and leave. Six months later, nobody on the team understood why anything worked the way it did. Because we hadn't built it. We'd just... accepted it."
"What would you do differently?"
"Track it. Every suggestion they made, we'd log whether we accepted it, rejected it, or modified it. And why." He paused. "Wait, so basically that's what the Creation Log is? An attribution ledger for ideas?"
"You tell me."
James thought about it. The collaboration log from Chapter 6 had tracked his interactions with AI. This was the same principle, but the stakes were different. In Chapter 6, he'd been analyzing a scenario someone else defined. Here, the solution was supposed to be his. If he couldn't trace which ideas originated in his head versus the AI's output, the word "his" didn't mean anything.
"Okay," he said. "I'll log every interaction. What I asked, what I got back, what I did with it."
"Good. Ten interactions minimum. And when you're done, I want you to answer one question: what was the single most valuable thing AI contributed, and what was the single most valuable thing you contributed?"
She picked up her notebook. "I'll check your log when you're finished. The log is the deliverable, not the draft."
Exercise 2: The AI Collaboration Draft
Layers Used: Layer 2 (Reasoning Receipt)
James is opening his Draft 1 alongside an AI tool, ready to collaborate. So are you.
You will use the Collaboration Log format from Chapter 6. Same structure, now applied to creative work instead of strategic analysis.
Build Your Draft 2
Using your Draft 1 as a starting point, collaborate with two different AI tools to develop your solution further. Maintain a Creation Log: every idea that came from you, every idea that came from AI, and every idea that emerged from the interaction that neither would have produced alone. This produces Draft 2.
Your Draft 2 (the improved solution after AI collaboration). The full Creation Log in table format: Interaction # | My Prompt | AI Response Summary | Decision (Accept/Reject/Modify/New Idea) | Attribution (Human/AI/Synthesis) | What I Added or Changed. The log must show at least 10 interactions. A brief note (100 words) identifying the single most valuable contribution AI made and the single most valuable contribution you made.
I am comparing my original unassisted solution (Draft 1) with my AI-collaborated version (Draft 2). I also have a Creation Log documenting every interaction.
Please: (1) Compare Draft 1 vs. Draft 2 -- is Draft 2 genuinely better or just longer/more polished? Identify the specific improvements. (2) Analyze my Creation Log: what percentage of the final solution's value came from me vs. AI vs. synthesis? (3) Are there places where AI's input actually made my solution worse (more generic, less original)? (4) Identify the elements in Draft 2 that are most original -- things AI would not produce if given only the original problem. (5) Rate my collaboration quality: was I leading the process or following AI's lead?
Draft 1:
Draft 2:
Creation Log:
Finally, complete the Thinking Score Card for this exercise: Independent Thinking (1-10), Critical Evaluation (1-10), Reasoning Depth (1-10), Originality (1-10), Self-Awareness (1-10). For each score, give a one-sentence justification.
Discuss with an AI. Question your scores.
Come back when you have your BEST evaluation.
Deliverable Template (click to expand)
CREATION LOG TEMPLATE
| Interaction # | My Prompt | AI Response Summary | Decision (Accept/Reject/Modify/New Idea) | Attribution (Human/AI/Synthesis) | What I Added or Changed |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 | |||||
| 4 | |||||
| 5 | |||||
| 6 | |||||
| 7 | |||||
| 8 | |||||
| 9 | |||||
| 10 |
Most valuable AI contribution: ___
Most valuable human contribution: ___
What Happened With James
James scrolled through his completed Creation Log. Ten interactions. Seven he'd marked as "Modify," two as "Accept," one as "Reject." The reject was an AI suggestion to restructure his entire approach around a stakeholder engagement matrix. It sounded professional, but it didn't solve the actual problem.
"I almost accepted that one," he said. "It sounded smart. But when I tried to explain it in my own words, I couldn't. That was the test: if I couldn't paraphrase it, I didn't understand it well enough to keep it."
Emma scanned the log. "Your most valuable AI contribution?"
"Interaction four. I asked the AI to poke holes in my implementation timeline. It flagged three dependencies I'd completely missed. I wouldn't have caught those on my own because I was too close to the design."
"And your most valuable contribution?"
"The framing. My original problem diagnosis from Draft 1 held up. The AI suggested a different angle twice and both times I came back to my framing because it was grounded in something I actually understood from my operations background."
"That's the pattern," Emma said. "You brought the framing. AI brought the stress test. Neither one alone would have produced Draft 2."
James looked at the attribution breakdown. Roughly 60% human-originated, 30% AI-originated, 10% synthesis. He'd expected the numbers to embarrass him. Instead, they felt honest.
The Lesson Learned
The Creation Log makes invisible collaboration visible. When you can trace every idea to its origin (human, AI, or synthesis), you stop wondering whether you contributed and start knowing exactly what you contributed. The paraphrase test is the simplest filter: if you cannot explain an AI suggestion in your own words, you do not understand it well enough to keep it.