Worked sample · printable PDF
POL-04 · Mock rulemaking
Run a mock comment period on a proposed AI school policy; produce the final rule and a comment/response summary that shows you actually changed something — or honestly explain why you did not.
Specialty centers · Algorithmic Policy & Governance
Two printable policy-and-governance briefs you can assign this term
Free leadership and social-sciences briefs about regulating a system you have actually taken apart. Students run a mock rulemaking with a comment/response record, and they practice a disparate-impact worksheet honest enough to refuse thin-cell overclaims. POL-04 and POL-02 have single-brief PDFs; every brief is in the bank PDF; the other ten are full on the page. No account. No student data leaves your school.
Ecological inference and governance humility
Report counts, not only rates, when subgroups are thin. Refuse causal language the design cannot support. Name the institution that would implement any recommended control, and give a rough cost. Public documents only — no insider data, no scraping behind logins.
How to run one
- Students choose the brief. You approve the draft policy (or system under study) and the stakeholder roles in writing first.
- Two scheduled checkpoints. Score with the six questions below.
- Mentor is advisory — school-supervised, never one-to-one unsupervised.
Featured packets
Featured · printable PDF
POL-02 · Disparate impact worksheet
Run a fairness-style audit on a public dataset with a metric set you declare first. Where samples are thin, refuse the claim rather than decorate it with a percentage.
Catalog (10 briefs on the page)
- POL-01 · Take apart a risk score
Using only public documentation, reverse-engineer how a scoring or ranking system claims to work — then list what you still cannot know. Mystery is a finding.
- POL-03 · Procurement questions that bite
Write the RFP questionnaire a division should send any AI vendor — mapped to themes in Virginia’s AI guidance as your teacher cites them from a current official source — not a vendor brochure.
- POL-05 · Treaty clause draft
Draft a narrow international clause on military or surveillance AI; red-team it for loopholes. Scope beats grandeur.
- POL-06 · Costed mitigation
Price one mitigation (human review, audit, or opt-out) for a hypothetical district of a size you state — show the arithmetic.
- POL-07 · Appeal without theater
Design an appeals process for an automated school decision that a student or family could actually complete in under two hours of effort.
- POL-08 · Shadow vs. production
Argue when a model should stay in shadow mode; define exit criteria with metrics the school could actually measure.
- POL-09 · Public records and models
Analyze what Virginia public-records logic implies for model documentation in a public agency — using public FOIA resources your teacher approves, not legal advice.
- POL-10 · Compare two governance frames
Contrast an EU-style risk-tier approach with a US sectoral approach on one concrete education or youth-services use case — without declaring a winner from aesthetics.
- POL-11 · Student data in the loop
Map an edtech AI feature against FERPA themes and your division’s published student-data rules; list data flows and red lines. This is a map, not a determination of compliance.
- POL-12 · Sunset clause
Write a policy that expires unless re-authorized with fresh evidence — and defend why sunsets belong in AI rules.
Six-dimension rubric
- R1 — Is one system, one institution, and one decision named?
- R2 — Could another student rebuild the analysis from the citations and worksheets alone?
- R3 — What happens under current policy — and does the proposal improve a named metric or right?
- R4 — Are thin samples, missing data, and enforcement gaps recorded?
- R5 — Could a school board or agency staffer act on the brief in one sitting?
- R6 — Who is harmed by a wrong rule — and what did the student refuse to claim?
Feedback
Scope unclear? Standards language off? Tell us here — you do not have to leave an email.