A Safe First Try

Subtitle: Try one bounded task

What you will make

Practice a small, fictional task while keeping the learner responsible for the input, the review, and any next decision. The goal is not to find a magic prompt. It is to notice what a tool can organize, what it cannot establish, and what a person must still decide.

Lab materials

• 📎 Resource · Maple Park source sheet — A fictional source and bounded request for the practice task.
• 📓 Notebook · Lab Notebook — An optional place to record the work and download a local copy.
• 📎 Resource · Learner guide
• 📎 Resource · A Safe First Try With Generative AI
• 📎 Resource · Maple Park checklist Gallery
• 📎 Resource · Facilitator guide

Before you start

Format: individual, pair, observation, or facilitated demonstration

Use only the 📎 Resource · fictional Maple Park announcement, a public source, or a facilitator-provided low-risk scenario. Do not paste private, client, medical, legal, financial, employment, account, credential, safety, or another person’s information. You may observe or use the supplied transcript instead of opening a chatbot.

The core idea

A useful result is still a draft. Keep the source safe, compare the output with what you supplied, and leave consequential decisions with a responsible person.

Try one bounded task

Complete the activity here. Open the optional 📓 Notebook · Lab Notebook beside this page when you want a place to record your work.

1. Make a bounded request

   Use the bounded request in the 📎 Resource · Maple Park source sheet exactly, or adapt it only for another safe fictional source.

2. Read the result as a draft, not an authority

   Compare the result with the source. A useful checklist might organize the time, meeting place, clothing, supplies, child supervision, and lightning condition. It should not invent a cancellation contact, equipment policy, or extra event details.

3. Do three human checks

   • What was useful?
   • What needs checking, correcting, or clarifying?
   • What decision still belongs to a person?

   Write one answer to each. If the tool invented a detail, move it to Questions to Ask or remove it; do not smooth it into a confident fact.

4. Set personal ground rules

   Record one low-risk use you may try, one type of output you will verify, one thing you will not paste, and one decision you will keep human-owned.

Check your result

• Compare your record with the source and the stated task. Name what the result supports, what remains uncertain, and what must be checked by a person or authoritative source before anyone acts.

• A polished result is not proof that it is accurate, complete, safe, or appropriate. Keep unresolved questions visible and verify consequential claims outside the tool.

You are done when

You are done when you can show the safe source or transcript, the request, three human checks, and four personal ground rules. The notebook can hold these notes locally and download them as Markdown or JSON.

• The safe source or supplied transcript
• The bounded request
• What was useful
• What needs checking, correcting, or clarifying
• The decision that remains human-owned
• Four personal ground rules
• One thing I deliberately did not share

Try another version

These are optional extensions, not additional completion requirements.

• Repeat with a fictional packing list or hobby plan
• Compare a useful but incomplete output with a carefully checked one
• Observe a facilitator instead of using a live tool
• Ask a partner to identify one detail that still needs verification