Session ownership audit

Mark every paragraph 'I generated' or 'AI generated' and count the ratio, so you can't argue with a number about who did the thinking.

When to use this: At the end of a session
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Checking readiness
How often to use it: Worth repeating often

When to use it

You are proud of a draft and suspect the reasoning spine is mostly AI.
A transcript review suggested low ownership and you want counts.
You need a number you cannot argue with about who did the thinking.

What to ask AI

Type something like: “Here is our transcript: [ ]. Do not re-answer content. Help me tag ownership (I / AI / mixed) line by line, then tally.”

Ask what ratio would be healthier for learning this skill, and one rule to get there.

Keep the hard part on your side: you accept the tally and change the next session’s rules.

How it helps

If the tool does the cognition, you don’t build it.

Ownership counts make invisible substitution visible.

Mark every paragraph I-generated or AI-generated, then look at the ratio.

Pride in the product with 90% AI tags on the reasoning is the failure mode this catches.

How to run it

Paste the work or transcript.
Tag I / AI / mixed without rewriting content.
Tally. Compare to a target ratio for this skill.
Set one rule for next time: attempt first, ask-only mode, or no draft generation.
Re-audit after the next session.

Example exchange

You

Here’s the transcript from my essay-drafting session. Don’t re-answer the content, help me tag who owned each part, line by line, then tally it.

Thesis paragraph: mostly AI. Example 1: mixed. Example 2: you. Conclusion: mostly AI. On reasoning-bearing lines, that’s about thirty percent you and seventy percent AI.

You

That’s lower than it felt. What’s a healthier ratio if I’m trying to learn how to argue?

For this skill, draft your claims first. Next rule: no paragraph generation until you paste your own outline bullets.

Copyable prompt

Here is our transcript: [ ].
Do not re-answer content. Help me tag ownership (I/AI/mixed) line by line, then
tally. Ask what ratio would be healthier for learning this skill and one rule to
get there.

The Tell

Here is how you know this method has flipped: pride in the product, and 90% AI tags on the reasoning.

You owned the glow, not the thinking.

Set one rule that raises your next tally before you polish further.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Productive struggle and desirable-difficulty arguments caution against over-scaffolding: if the tool does the cognition, the learner doesn’t build it. Counting ownership makes substitution visible. Principle moderate for ownership-as-metacognition; AI paragraph audits are delivery-speculative.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

Automated ownership tagging can err or flatter. Delivery speculative. Trust the tally only after you agree with the labels.