Optional Studio: Showcase

The optional studio is the showcase. There are no tools and no new concepts. Participants share what they built, found, or made across the three sessions. The session closes with a collective reflection and a short closing ritual.

What participants bring

By the end of this session participants will have:

• Presented their final project in 8–10 minutes
• Heard and responded to 4–6 other projects
• Named one thing they are taking back to their practice
• Identified one open question they are still holding

Projects can come from any pathway:

• Make — a creative artifact with evidence of iteration (at least three versions)
• Teach — a classroom-ready activity, worksheet, or facilitation plan
• Investigate — a model behavior experiment or Default Test with a written claim
• Explain — a tool, poster, zine, or concept bridge
• Critical / No-AI — a critique, consent checklist, model card, or unplugged activity

Session flow

Opening (0:00–0:10)

Welcome to the optional studio. This is the session where the camp becomes yours.

Over the last three sessions we asked the same question in three different registers: what is the machine actually doing? We saw it predict tokens, draw defaults, drift across frames. We named what was human, what was machine, and what was assumed by the system.

Today you show what you found.

A few norms for the share-out:

• We respond to evidence, not to impressiveness.
• “That surprised me” and “I don’t know what to make of that” are complete responses.
• We name what the human decided.

Participants have 8–10 minutes. There is no required format. Show what you made, explain what you found, or describe what you would teach. Then the group asks two questions.

Presentations (0:10–1:25)

Each participant presents for 8–10 minutes, followed by 2 minutes of response. Facilitator keeps time. After each presentation, ask the group two questions:

• “What did you notice?” — factual observation, not evaluation
• “What question does this open?” — something the project raised that wasn’t settled

Do not open the floor to general feedback. The two-question format keeps responses substantive and prevents the session from drifting into praise or critique.

Consider ordering presentations so that different pathways are distributed across the session — an investigator next to a maker next to a teacher.

Collective reflection (1:25–1:50)

Round 1: One thing you are taking back (5 min)
Go around the group. Each person completes: “I am taking back ___.” No follow-up, no elaboration.

Round 2: One question you are still holding (5 min)
Same structure. Each person completes: “I am still wondering ___.” These become the shared open archive of the cohort.

Round 3: What the machine can’t do (10 min)
Open discussion. Prompt: “Based on what you saw this summer: what is something the machine genuinely cannot do that you think matters?” Push for specificity — not “understand,” but what understanding would require that prediction doesn’t.

Closing (1:50–2:00)

What we did this summer was investigate. Not evaluate, not celebrate, not panic — investigate.

You built a protocol: hypothesis, controlled test, output comparison, evidence-based claim, ethical reflection. That protocol is portable. It works on any model, any modality, any tool that shows up in your school or your community or your studio.

The repeated question — what is the machine actually doing? — does not have a final answer. The systems change. The answers change. But the practice of asking, with evidence, with specificity, with attention to whose labor and whose images and whose stories shaped the training data — that practice is yours now.

Collect from each participant: a one-sentence project description, credit preference, permission to link in the recap, and any follow-up interest. Use the AI Use + Consent Checklist before publishing anything.

Facilitation notes

• If a participant says “I didn’t finish” or “I’m not sure this counts,” respond: “Show us what you have. What you found in the process is the project.”
• Watch for participants who describe their project but never show it. Prompt: “Can you show us the actual output / artifact / document?”
• If a participant used no AI tools, their project is valid and often the most interesting in the debrief.
• If presentations run long, keep the two-question format strict. Let one presentation run to 15 minutes — but pull time from collective reflection, not from other presentations.
• If presentations run short, use the extra time in collective reflection. The “what the machine can’t do” discussion often goes longer than expected.