Facilitator guide: It Starts With a Question

Labs reference: Facilitator guide: It Starts With a Question.

Purpose

Help learners use live generative AI to follow one genuine question without turning curiosity into a race for a complete answer. The learner owns the question, the AI job, the supporting methods, the record, and the decision to continue or stop.

Suggested session

Use demonstration → learner practice → reflection for Zoom and in-person sessions. Demonstrate one prepared inquiry first. Then let learners work individually, in pairs, or in groups on a question they choose or adopt. A group may share a question and divide roles; the learners make the exploration, inspection, and next-step decisions. They may direct a partner or facilitator operating the chatbot.

Watching the opening example alone is preparation, not completion of the lab. If a shorter session ends after the example, name the learner practice still to do. Self-study learners may inspect either prepared example before following their own question; the completion criteria are the same.

For provisional planning only, the existing learner-practice segments total 33–38 minutes: 5 to notice and shape a question, 5 to select an AI job, 10–15 to explore, 8 to sort, and 5 to choose a next step. Allow separately for the opening demonstration, voting, setup, transitions, and discussion. Refine these estimates from practice; they are not learner deadlines.

Pilot within the Curiosity track

For a first Curiosity track pilot, follow the short kickoff in Becoming Curious Again: let learners choose a question, use the capability table to choose a relevant AI job and tool, demonstrate one opening example, then have learners follow this lab individually, in pairs, or in groups. Use the module’s deeper discussion for optional preparation or follow-up. Begin with this compact path before expanding the route backlog.

Gather track-level feedback about the promise, starting point, route clarity, status labels, and missing or misplaced modules and labs. Use the evidence-of-learning prompts below to hear where a useful result came from, what supports it or still needs checking, and the learner’s next step or reason to stop. Automated checks and facilitator impressions do not replace learner evidence.

Facilitator demonstrations can help while learner project examples are missing. The learning continuation has two prepared study walkthroughs, creative play has two prepared making walkthroughs, and simulation has two prepared model walkthroughs. Read both examples for the chosen route before offering a class vote. After the demonstration, learners make their own choices individually, in pairs, or in groups. Learner presentations are optional.

Preparation and participation

Provide several low-risk objects, observations, images, claims, or everyday puzzles for learners who do not arrive with a question.
Offer spoken, written, drawn, paired, observer, timekeeper, source-finder, and facilitator-led routes.
Prepare one approved live chatbot and review both available opening walkthroughs before offering the class a vote. Check the source links and the tool capabilities needed for either choice. Model the selected inquiry before learner practice.
The Gallery offers two written opening walkthroughs: refrigeration and tallest mountain, each with source links and prompts to run live. At the opening of the session, let the class vote on the question, then choose a branch. Briefly explain that these are prepared teaching examples, not recorded learner sessions. Open one source together and revise one claim.
Open the Lab Notebook and Question Path before the session so you can explain the options. The Notebook carries the complete activity one step at a time; the Question Path offers a printable or separate record. Both are optional, and learners do not need both. Spoken, drawn, and group participation remain valid. Notebook entries are not submitted or saved automatically; learners can export a local copy if they want to keep their work.
Match the tool to the learner’s intended use: conversation, searching sources, creating a draft or image, or exploring a stated scenario. Check that the available tool has the capability you intend to demonstrate. One useful job is enough; learners do not need a tour of every tool.
Do not assign question difficulty, tools, or roles by age or generation.
Adjust depth, vocabulary, pacing, and written load to the group without assuming ability or technology comfort from age.
Allow learners to pass, change questions, use invented examples, or keep their Question Path private.
Do not require an account, recording, personal story, polished artifact, or public presentation.
Let each learner operate the tool, direct a partner, or guide and observe the facilitator. If live AI is unavailable, help learners prepare their question and AI job, but do not count the core activity as complete.

Facilitation priorities

Protect genuine learner choice. Help narrow an unworkable question without replacing it with the facilitator’s question.
Ask “What happened?” before “What does it mean?” so observation and interpretation remain distinguishable.
Treat uncertainty and new questions as results, not failures to finish.
Keep AI’s job proportionate to the question. Use it for one named purpose rather than giving it open-ended authority.
Ask two separate questions about an important result: “Where did this come from?” and “What supports it, or still needs checking?” Help learners distinguish observation, source reports, generated material, and interpretation in ordinary language. Uncertainty can apply to any of them. A second generated response is not independent verification.
Invite observation, a small experiment, a source, a knowledgeable person, or another method wherever it strengthens the exploration.
Celebrate a reasoned stopping decision.

Safety and access

Use public, fictional, learner-owned, or low-risk material. Do not ask learners to disclose private information or investigate consequential medical, legal, financial, employment, credential, or personal-safety questions in this activity. Stop an unsafe experiment. For an important factual claim, help the learner locate an independent source or relevant human expertise.

Read prompts aloud when useful, enlarge source text, allow additional processing time, and accept movement, drawing, pointing, dictation, or conversation in place of writing. The optional Question Path should reduce cognitive load rather than become a compliance form.

Evidence of learning

Look for the learner’s ability to:

state or represent one question they care about;
select an AI job that fits the question and explain why;
participate directly, through a partner, or through a facilitator-operated demonstration;
explain where an important result came from and what supports it or remains uncertain;
name what they would verify before repeating or acting on a claim; and
choose a next step or a reason to stop.

Productive outcomes include revising the question, discovering that the first method was a poor fit, finding conflicting sources, or deciding that the available evidence is not enough. Do not score the learner on topic sophistication, volume of notes, technology use, or whether the question was answered.

Debrief prompts

What changed in your question?
Which method helped, and what could it not show?
What did you observe directly?
What came from another person, source, or tool?
What are you choosing to leave uncertain?
What would make another step worthwhile?

Provisional timing for facilitator planning

These inherited estimates are starting points for facilitation, not completion targets for learners. Refine them after observing setup, access needs, discussion, and actual participation.

Earlier lab estimate: 30–45 minutes. The segment breakdown above is a planning starting point; it excludes the opening demonstration and other session overhead.

Source: labs/it-starts-with-a-question/facilitator-guide.md