Banyan Global Learning
AI Policy hubFree checklist

District AI Policy Checklist

Twenty-six questions your AI policy — and the governance around it — needs to answer. Organized by the five layers, because a policy that only covers compliance fails everywhere else. Bring it to your next cabinet meeting and mark what you can honestly check.

Layer 1 · Strategic Clarity

Before policy language: does the district know what it believes?

  • A board-approved statement of AI intent exists — one page, plain language, dated within the last 18 months.
  • The superintendent and cabinet can state the district’s AI posture in two minutes without a script.
  • AI strategy lives inside the district strategic plan, not in a separate binder nobody opens.
  • Success metrics are defined for both efficiency (staff time, cost) and learning quality — not just one.
  • A parent or reporter can find the district’s AI position on the public website in under a minute.

Layer 2 · Risk & Compliance

The layer most policies cover — but usually only halfway.

  • An inventory of every AI tool in district use exists and has been updated in the last 90 days.
  • Vendor evaluation includes AI-specific criteria: training data, bias, data flow, transparency.
  • The acceptable use policy names AI explicitly and was revised within the last 12 months.
  • FERPA and COPPA implications of each approved tool have been reviewed and documented.
  • Staff know the incident protocol for a data breach, a biased output, or an integrity violation — by name, not by guess.
  • Academic-integrity boundaries for AI are differentiated by grade band, not one rule K–12.

Layer 3 · Instructional Integration

The cognitive core — the layer that determines whether AI helps or hollows out learning.

  • The district has articulated which cognitive work must remain human-performed for learning to occur.
  • High-stakes assessments have been reviewed against what AI can now do, and updated where they broke.
  • Teachers have a shared framework for deciding when AI assists learning versus when it bypasses it.
  • Professional development covers pedagogy, ethics, and cognitive impact — not just tool walkthroughs.
  • Students are taught to evaluate AI outputs: what to trust, what to verify, what to reject.

Layer 4 · Operational Implementation

The tissue between the policy document and Monday morning.

  • A 6–12 month AI roadmap exists with milestones and named owners.
  • PD is sequenced over time and differentiated by role — not a one-and-done workshop.
  • A proactive parent and community engagement plan exists (not just reactive FAQs).
  • Students have a structured voice in AI governance decisions that affect them.
  • Building leaders apply consistent expectations across schools — checked, not assumed.

Layer 5 · Infrastructure & Future-Readiness

The decisions that quietly lock a district in — or keep it free.

  • A decision framework exists for cloud versus local/private AI solutions.
  • Infrastructure choices trace to educational principles, not vendor sales cycles.
  • Procurement criteria require interoperability and data portability.
  • Someone owns monitoring what’s next: agentic AI, multimodal models, AI companions.
  • A 2–3 year evolution roadmap exists that assumes the technology will keep moving.

How many could you honestly check?

The checklist tells you what to look at. The free AI Snapshot tells you where you stand: about six minutes, a score for each layer and up to three next steps.

Take the free Snapshot