We can explain why AI was or was not appropriate.
Build a game with your class
Two human-only moves.
Three student-choice decisions.
Goal and Audience stay human-only. At Mechanic, Engagement, and Show it, choose AI or no AI, record why, and treat both pathways as equally valued.
G · Human-onlyA · Human-onlyM · Student choiceE · Choice after real playS · Student choice
Teacher setup
Set the guardrails before students design.
GHuman-only
Move 01
Goal
What should players try to accomplish—and what should they understand because of it?
Why no AI? People decide what is worth learning, what responsible success means, and which boundaries matter.
AHuman-only
Move 02
Audience
Who will really play—and what will help them enter, participate, and influence the design?
Why no AI? AI cannot represent classmates, lived experience, cultural knowledge, disability, or belonging. Ask people.
MStudent choice
Move 03 · Student-choice AI
Mechanic
What repeatable player action will make the target thinking playable?
Human firstWrite the target thinking, audience needs, materials, and boundaries before opening an AI tool.
Our choice for this move
Optional prompt frame
Generate three contrasting, low-cost game loops that require players to [target thinking]. Players are [audience]. Respect these constraints: [limits]. Do not write a finished game or invent subject-area facts.
No-AI equivalent: Each team sketches three loops in six minutes, then swaps with another team for critique.
EStudent choice
Move 04 · Student-choice AI after real play
Engagement
What do real players do, think, feel, and need while they play?
Watch forEvidence observedRevision decision
Our choice after real playtesting
Optional prompt frame—after real play only
Using only these de-identified observation notes, group recurring observations and suggest two patterns we should investigate. Flag uncertainty. Do not simulate players, judge belonging, or make a revision for us.
No-AI equivalent: Affinity-map the observation notes with sticky notes, name possible patterns, and test each one against the original evidence.
Non-negotiable: Only real players can show whether the game creates agency, confusion, exclusion, curiosity, or belonging. AI may sort notes; it cannot replace that evidence or the student interpretation.
SStudent choice
Move 05 · Student-choice AI
Show it
What proves the game works—and what can each student demonstrate beyond the game?
Human firstFinish the prototype, learning claim, and evidence record before asking AI to challenge the work.
Our choice for this move
Optional prompt frame
Act as a skeptical reviewer. Identify unclear rules, unsupported claims, missing perspectives, and places where this mechanic may not demonstrate [learning goal]. Ask questions. Do not rewrite our work or add new facts.
No-AI equivalent: A peer team acts as the skeptical reviewer using the same classification and verification process.
AI literacy evidence
Can students explain their judgment?
AI use itself earns no credit. Evidence comes from the quality of the human decisions surrounding it.
We identified limitations, bias, errors, and missing perspectives.
We checked claims against reliable sources and real playtests.
We recorded prompts, outputs considered, changes, and attribution.
Each student can perform the target thinking without AI.