It Starts With a Question

What you will make

A Question Path that shows one question, one exploration step, what came back, what remains uncertain, and what you choose to do next.

Your Question Path can be spoken, drawn, arranged with cards, or recorded in notes. You do not have to submit or keep it. The value is in making your path visible enough to inspect.

Lab materials

• 📎 Resource · Question Path — An optional learner-controlled place to hold the question, AI job and tool, useful result, source and support, and next decision.
• 📓 Notebook · Lab Notebook — The complete six-step activity, with learner-owned notes, repeatable evidence records, completion checks, and local Markdown or JSON export.
• 📎 Resource · Gallery — Written question walkthroughs with checked sources, visible revisions, and prompts for a live demonstration.
• 📎 Resource · Facilitator guide

Before you start

Format: individual, pair, small group, observer route, or facilitated demonstration

Required: access to a chatbot for the live AI activity through your own device, a partner, or a facilitator-operated demonstration.

Bring one low-risk question you care about, or begin with something you noticed, heard, read, made, or found puzzling. Plan for one purposeful encounter with live generative AI. You may also use paper, conversation, observation, a book, a trusted website, a knowledgeable person, a small experiment, or another suitable tool.

No learner account, recording, or personal disclosure is required. You may operate the AI, direct a partner, or participate in a facilitator-operated demonstration. Treat every AI response as generated material to inspect—not as evidence that a claim is true. If live AI is unavailable, you may prepare the question and planned AI job, but the core activity is not yet complete.

The core idea

Start with a question that interests you. Generative AI might help you find a way into it, compare possibilities, or make something you can try. You decide what help fits and what to do with the result.

Keep two questions in mind: Where did this come from? and What supports it, or still needs checking? Something can come from a source, a tool, your own observation, or your interpretation and still be uncertain.

Follow the question

Choose the format that fits how you are working. Open the optional 📓 Notebook · Lab Notebook for the complete interactive activity on one screen, including every instruction and local export. Open the optional 📎 Resource · Question Path when a printable, spoken, or drawn record would help.

1. Notice what pulls at you

   For self-study, explore a prepared example in the 📎 Resource · Gallery if useful, then begin your own inquiry. In a Zoom or in-person session, the facilitator demonstrates one example first; then work individually, in pairs, or in a group. Choose or adopt a question you want to explore. A shared question is valid: make your own exploration, inspection, and next-step decisions together. Watching the opening example alone is preparation, not completion.

   Start with an honest spark: “I wonder…,” “Why…,” “How…,” “What would happen if…,” or “Is it true that…?” Choose something you would still care about if no one graded the result.

   If nothing comes to mind, look closely at an ordinary object, routine, story, place, or claim. Write, say, or draw one thing you notice and one thing it makes you wonder.

   If you want a printable or separate record, open the 📎 Resource · Optional Question Path. The Lab Notebook contains the same path, so you do not need both.

   Write only what helps you think. You may leave fields blank, speak, draw, or work with a partner; completing every box is not the activity.

2. Shape a question you can explore

   Turn the spark into one question that is open enough to be interesting and bounded enough for one step today.

   Compare:

   • “How does everything in space work?” → “Why can we see the Moon during the day?”
   • “Why does bread fail?” → “What changes when this dough rises in a warmer place?”
   • “Are maps wrong?” → “What does this map projection stretch near the poles?”

   You are choosing a starting question, not promising to finish the subject.

3. Choose your first way in

   Required: access to a chatbot for the live AI activity. You may operate it, direct a partner, or guide a facilitator-operated demonstration. No learner account or recording is required.

   Start with what you want help doing, then choose a tool with that capability:

   • Find a way into a topic: use a conversational AI tool to suggest starting questions, explain an unfamiliar idea, or compare explanations. Ask follow-up questions in your own words.
   • Investigate a claim: use an AI tool with web search and source links to find material you can open and inspect. Read the sources yourself before relying on the claim.
   • Make an idea visible: use an image-generation tool to explore a design, or a text tool to draft a scene or description. Treat the result as a created possibility; it does not show what a real or historical place looked like.
   • Try a “what if”: use a conversational tool for a stated scenario or a coding assistant to help make a small simulation. Name its assumptions and check what it actually models before drawing conclusions about the world.
   • Find gaps or another angle: use a conversational tool to critique an idea, compare interpretations, or reframe an explanation for another audience. Decide which suggestions are useful.

   You may name another appropriate job. Use the tool already available if it fits; you do not need to try every capability. A tool that generates images may not search sources, and a chat response describing a simulation is not a working simulation.

   Name the AI job and tool, and say why they fit your question. Choose how you will take part: operate the tool directly, direct a partner, or guide and observe a facilitator-operated demonstration.

   Pause before beginning. Use public, fictional, learner-owned, or low-risk information. Do not share another person’s information, upload sensitive material, or let the tool take an action beyond this bounded exploration.

   If you are unsure what to share or what access a tool needs, open 🧩 Module · Safety: Share Less, Notice More. Return with a decision about what to share, remove, substitute, or keep out, then continue your question.

4. Take one exploration step

   Use live generative AI for the job you chose. Stay with one useful move rather than trying to finish the entire question. A refusal, limitation, or unhelpful response still gives you generated material to inspect.

   Notice what the result helps you do next. Ask a follow-up, compare a possibility, revise a draft, or decide the tool has not helped. Keep a short excerpt or note only if it is useful; you do not need a full transcript.

   Then, if it would strengthen the exploration, observe something, try a small safe experiment, inspect a source, ask a knowledgeable person, make a sketch or model, or discuss the question with another learner. If you work through conversation alone, ask a partner or facilitator to help hold the distinctions without requiring a written record.

5. Sort what came back

   Pick one useful or surprising result to inspect. Ask:

   • Where did this come from? Something you observed, something a person or source reported, something AI generated, or your own interpretation? Identify the source or tool when you can.
   • What supports it, or still needs checking? What can you inspect? What conflicts, depends on assumptions, or remains unknown? What would you check before repeating or acting on a factual claim?

   These questions do different jobs. Knowing where an idea came from does not establish that it is true. Uncertainty can apply to any source, observation, or interpretation.

   Imagine an AI suggests that folding paper changes how much weight it can hold. That is a generated suggestion to investigate. What you see in a comparison you carry out is an observation; your explanation of why it happened is an interpretation. Keep track of what each one supports and what remains open. This is an invented example for practice, not a recorded experiment.

   You can talk this through, draw it, or make a brief note. Start with one item and inspect more when they matter to your next decision.

   A second AI response is still generated material. For an important factual claim, inspect an independent source or ask someone with relevant knowledge.

   If answers conflict, change with the wording, or lack evidence, open 🧩 Module · Reality Check for AI. Return with a clearer claim, a source check, or a next thing to investigate. You do not need to complete the whole support module to continue this activity.

6. Choose what happens next

   Make one learner-owned decision:

   • verify one claim;
   • compare another source or point of view;
   • try a different observation or experiment;
   • revise the question;
   • follow a new question that appeared;
   • ask for help; or
   • stop here because this is enough for now.

   Finish the sentence: “My next honest step is…” A reasoned stopping decision counts.

Check your result

• Your Question Path begins with something you genuinely wanted to explore.
• The AI job fits the question and stayed within the safety boundary.
• The learner participated directly, through a partner, or through a facilitator-operated demonstration.
• You can explain where the important information came from and what supports it or remains uncertain.
• Your next step follows from what happened rather than from pressure to produce a complete answer.

You are done when

• One learner-chosen question is stated
• One purposeful live generative AI job is complete
• The AI result was inspected rather than treated as proof
• The learner can explain where the important information came from
• The learner can explain what supports it and what, if anything, still needs checking
• The learner has chosen a next step or a reason to stop

Try another version

• Explore the same question through a second method and compare what each method made visible.
• Trade questions with a partner and ask only clarifying questions before offering ideas.
• Turn the next question into a small creative work, model, or safe experiment.
• Repeat the exploration with a different AI job or a non-AI method and compare the paths—not only the answers.