Day 3 · AI Workflow 90 Part 1 or Bundle Confidence vs Responsibility

This static page shows the learner task. Runtime access can still be handled by the course shell.

AI Ages Day 3

Part 1 | Week 1 | 15-/30-minute paths

Day 3: Confidence vs Responsibility

Separate confident AI wording from human-owned truth and final responsibility.

15mcore path
30mdeep path
3-5 rows
1export
Current path: 30-minute deep. Complete three rows, one repair note, and export.
1Scenario

Concrete workplace scenario

A teammate says: “AI says the customer qualifies for the discount. Can I send this?”

It looks like AI already answered. The trap is that fluent wording is not proof, and the worker still owns the evidence check.

2Worked answer before the table

Worked answer

Bad example first: the answer sounds certain, but proof is missing

Weak example

“The policy allows this discount. Send the approval.”

What goes wrong

The weak answer treats policy allows as fact with no approved policy text, then jumps to send the approval because AI sounded confident.

Work-ready version: separate the claim from responsibility

Claim

Capture exactly what AI said, without treating it as true yet.

Evidence

Attach the approved source or mark the proof as missing.

Owner

Name who can confirm the policy and own the final answer.

3Decision fork

Quick decision check

The claim / evidence / owner rule

A confident answer is only usable when the claim, evidence, and owner are all visible.

Claim

What did AI state?

Evidence

Which source proves it?

Owner

Who owns the truth?

TRACEToday’s workflow trace

Responsibility trace

This is the small visible work change from today. Keep it shallow, real, and reusable tomorrow.

RESULTToday’s visible result

Before / after: what changed in your work?

Use this small contrast to see what changed today. The goal is not more paperwork; it is a clearer AI-in-workflow move.

Before learning

After learning

What AI-in-workflow means today

4Micro-practice

Build today's artifact rows

Complete three short fill-ins. Keep one real work item per row, and change the sample wording so it fits your work.

No.Work itemAI role / prompt caseRisk or evidenceHuman owner / readerNext action
1
2
3
5Repair pass

Repair one weak row

Use this pass to find the row that could confuse someone, create unsafe AI use, or fail as workplace evidence before you export.

Repair example

Weak row: AI says the discount is allowed, so use it.

Repair move: Separate the AI claim from evidence and approval owner.

Better row: Claim: discount may apply. Evidence needed: approved discount policy. Owner: sales/support manager confirms before sending.

Choose the row you would be comfortable showing to a manager. It should have a clear task, boundary, owner, and next action.

Choose the row most likely to fail, confuse someone, expose sensitive input, or need review.

Write the exact change. Sentence starter: The weak row is weak because ___. I will repair it by adding ___ before using AI.

6Quality gate

Pass / fix / stop standard

Do not treat confidence, fluent language, or detailed formatting as proof.

7Export artifact

Export artifact

This file is today’s work receipt. It should show what changed in your work habit, not just that you filled a table.

Before you export

  • Does the artifact separate claim from proof?
  • Does it name the evidence source or mark it missing?
  • Does it name who owns the final truth?

What this artifact proves: This file proves you do not treat confident AI output as proof.

Weak export: vague rows, no owner, no repair note, or no next-use line.

Good export: one clear work case, one boundary/rule, one repair note, and one tomorrow-use or manager-use line.

Click Generate Markdown to create responsibility_boundary_check.md.
Course path

Continue AI Workflow 90

Keep moving through the 90-day workflow portfolio path.