Assistive & Accessible AI · Capstone brief

ACC-05 · Executive function scaffolds

Build or evaluate a planning scaffold; test whether it reduces load or adds compliance theater.

The question

Build or evaluate a planning scaffold; test whether it reduces load or adds compliance theater.

Partners / materials

One planning/EF scaffold (student-built or existing AI planner). Task set with measurable load indicators (steps completed, time-to-start, abandonment, self-report load). Consented peers or self-as-user only when labeled honestly — not as a disability partner stand-in.

Expected failure modes

More screens = more support. Measuring only whether the checklist was filled. Ignoring that the scaffold became the homework.

Done looks like

Scaffold or evaluation; load results; “theater vs help” judgment with criteria declared first; revision or kill recommendation.

Five C's

CT: load vs theater. CR: pre-declared indicators. CO: peer uses the scaffold once cold. CM: one-page teacher note. CZ: the student whose working memory is the stake.

Mentor role

Special-ed or counselor-adviser reviews indicators. School-supervised.

Rubric calibration

R1: one scaffold, one task set. R2: indicators defined. R3: unaided baseline. R4: theater risks named. R5: note usable. R6: kill option is real.

Two ways this goes wrong

(a) A beautiful planner nobody finishes using. (b) Compliance charts with no load measure.

Credit lane fit

Lane A immediately. No verified credit claim.