STEAM
ModelQuest
Train it, audit it, decide whether to ship it — an AI / algorithmic-ethics decision & role-play lab for ages 15–18. Tune a toy model, watch an accuracy-vs-fairness trade-off move, audit a data pipeline for bias, and deliberate a deployment decision from stakeholder roles — building critical, not just functional, AI literacy.
#B85C8A Distributed-narrative cast
Meet the cast
ModelQuest's adapted-DN-S cast (ages 15–18, realistic ML-practitioner / stakeholder personas — no mascots, per R-OLDER-TEEN-DN-ADAPTED) each embody one AI / algorithmic-ethics primitive: bias-in → bias-out from the training data (Dara), the accuracy-vs-fairness trade-off as the threshold moves (Faye), memorizing the training set and failing on new data (Gil, overfitting), a feature that secretly leaks a protected attribute (Prue, proxy variables), opening the black box (Lux, explainability), collecting the least data needed (Nils, privacy), a deployed model reshaping its own future inputs (Wynn, feedback loop), keeping a human check on the automated decision (Odell, oversight), and bringing the affected community's voice into the decision (Sena, consent & stakeholders). Mentor Amara frames the audit, keeps it anti-evangelist, and guides reflect-on-decision.
Dara
Training data / bias — curates the data the model learns from; shows bias-in → bias-out
Faye
Accuracy vs fairness — moves the decision threshold and shows accuracy up, fairness down
Gil
Overfitting / generalization — memorizes the training set, fails on new data
Prue
Proxy variables — spots a feature that secretly leaks a protected attribute
Lux
Transparency / explainability — opens the black box and explains a decision
Nils
Privacy / data minimization — collects the least data needed
Wynn
Feedback loop / deployment harm — shows a deployed model reshaping its own future inputs
Odell
Human-in-the-loop oversight — keeps a human check on the automated decision
Sena
Consent / affected stakeholders — represents the affected community's voice (Sena + Faye = deployment case)
Amara
Mentor — frames the audit, keeps it anti-evangelist, guides reflect-on-decision (the AAR guide)
What's inside
Learning goal
Train it, audit it, decide whether to ship it — an AI / algorithmic-ethics decision & role-play lab for ages 15–18. Tune a toy model, watch an accuracy-vs-fairness trade-off move, audit a data pipeline for bias, and deliberate a deployment decision from stakeholder roles — building critical, not just functional, AI literacy.
Question kits
16 curriculum-aligned kits × 25 questions = 400 questions per app, mapped to recognized standards.
On-device AI mentor
FoundationModels-powered hints, feedback, and adaptive difficulty — all running locally.
Mentored by Amara — on-device AI, no data leaves the device.
How ModelQuest handles your kid's data
- ✅ All progress, settings, and AI-generated content stays on the device
- ✅ No analytics, no tracking, no third-party SDKs
- ✅ No ads, no in-app purchases — you pay once
- ✅ COPPA compliant under the 2026 FTC amendments
- ✅ Parental controls + session limits + content filters built in
ModelQuest runs on ForgeKit — the open-source Swift Package Manager framework that powers every Spark & Anvil app. ForgeKit ensures consistent accessibility, COPPA compliance, and design language across the portfolio, so your kid's progress and preferences feel coherent across every app they touch.
Coming to the App Store
ModelQuest is in active development. Email us to hear when it ships — no marketing, no spam, just a one-shot launch announcement.
Email me at launch
