ParalleX LabsBefore You Paste

Facilitator guide: Before You Paste (20 to 30 minutes)

This is a browser practice lab and human-review training environment. It is not a privacy protection tool, an AI fact checker or a safe translator. The software never shows a green "safe", "verified" or "approved" badge.

Learning objective

Participants practise identifying personal and sensitive details in synthetic humanitarian text, answer permission questions before scanning, decide what to replace or generalize, and understand that this lab is a last check, not permission to use prohibited data.

Maps to kit Session 1

Aligns with Humanitarian AI Training Kit Session 1 (responsible foundations): data responsibility, protection lens, and organizational rules before tool use. Companion: Humanitarian AI Risk Screen.

Timing

Activity run sheet
TimeActivityNotes
0:00 to 0:03Welcome, core message on home screenEmphasize: allowed data only; redaction does not change prohibition
0:03 to 0:08Demo: pattern check on short synthetic noteNo AI account needed; works offline after load
0:05 to 0:08Permission questions demoShow what happens when tool or purpose is not approved
0:08 to 0:20Practice mode: one EN and one FR passage, plus wrong-use scenarioParticipants mark items, then reveal answer key
0:20 to 0:25DebriefSee questions below
0:25 to 0:30Optional: optional NER downloadSkip on low bandwidth; pattern layer is enough for learning

Debrief questions

  1. What would your organization's policy say about the practice passage you reviewed?
  2. Which findings would you replace versus generalize versus keep? Why?
  3. Where could combined details identify someone without a full name?
  4. What is one habit you will adopt before pasting into an AI chat tool?

Low bandwidth and no AI account

The lab runs entirely in the browser. Pattern and context checks need no download. Optional name recognition (~129 MB) can be skipped. No generative AI login is required for the exercise.

Accessibility notes