← All work

Compliance · Branching micro-course

Call It Right

An AI governance micro-course for Texas public school staff. Five real decisions, three possible verdicts, and no slides of statute.

My role
Instructional designer and researcher
Audience
Texas K-12 teachers and instructional staff
Format
Interactive HTML scenarios + storyboard
Status
Prototype complete · animated version in pre-production

The problem

Teachers already use AI to draft passages, screen readers, and take notes. Texas now has rules that touch that work, including TRAIGA (HB 149, effective January 1, 2026). Most compliance training reads the law aloud and ends with a quiz on definitions. Staff pass it and still don't know what to do at 7:50 a.m. with an AI-made worksheet in hand.

Performance goal: staff verify AI output against the TEKS and their own evidence, make consequential decisions themselves, and tell families plainly when a tool helped.

What I did

  1. Checked the law before writing a line. Research corrected two common claims: the rule on confidential-meeting notetaker bots comes from state prohibited-technology policy, not the education agency, and TEKS review is a teacher-of-record duty under district policy, not a standalone AI statute. The course says so, and points staff to district counsel.
  2. Action-mapped five moments where staff actually meet AI: materials before the bell, a screener's auto-drafted note to a family, and more.
  3. Built the branching course with consequence feedback for every choice.
  4. Wrote the five-module storyboard and script book, with a locked character reference for the protagonist, Ms. Dawson, and a production plan for animated, lip-synced scenes.

Key decisions

Three verdicts, not two

Every scenario has a violation, a right call, and an overcorrection. Tossing the AI draft and hand-writing everything is compliant but wastes the time the tool saves. The goal is verify, then own it.

Practice in a new skin

In the script book, each behavior comes back in a different situation after its scenario. Example: the group has learned closed syllables only, the AI made three "decodable" words, and the teacher taps the one that isn't ready yet.

Honest attribution builds trust

A compliance course that cites the wrong rule is a liability. Every verdict chip names the real source, and the closing screen says to verify specifics with district counsel.

What's next

The animated version pairs wide establishing shots with lip-synced close-ups of Ms. Dawson, using one locked character reference across tools so she looks the same in every scene. A short follow-up check at 30 days would show whether staff are disclosing tool use in practice.

Make the calls yourself

Five decisions. Pick what you would really do on a busy school day.