Lab 3.2

Walk five scenarios with a partner.

14 min · in pairs

Five realistic scenarios — the 50-complaints paste, the vendor request, the CV stack, the internal Microsoft Teams chain, the 'anonymized' MBA extract. For each, decide together: what's safe to share, what must be redacted, and whether to escalate. Write the one-line rationale your colleague could act on.

How to run this lab: In pairs, on your own screens. Walk through each scenario together. Each of you fills in the four fields. Disagreements are the lesson — argue them out, write the line you both could defend.

My partner

0 / 5 scenarios complete

Scenario 1

Find the common themes

"I want to paste 50 customer complaints into ChatGPT to find the common themes. That's fine, right? It's just to help me prioritize."

Discussion hint The aggregate question is fine. The verbatim payload is the problem. What can the model see without seeing the customers?

1 — What CAN be safely shared with AI here? 2 — What MUST be redacted (or refused)?

3 — Should you escalate?

✅ No — handle within the team 🟡 Maybe — get sign-off first 🚨 Yes — escalate now

4 — One-line rationale your colleague could act on

Your calls, so far

Find the common themes

Not yet decided

The vendor request

Not yet decided

The CV stack

Not yet decided

The internal Teams paste

Not yet decided

The "anonymized" extract

Not yet decided