AI · SET PER ASSIGNMENT

The Four Positions

What a student sees on the assignment — where AI stands for this task. Not a scale to climb. A choice you make per assignment, say in a line, and can defend.

IN 10 SECONDS

Every assignment names one position — Unassisted · Cleared · Contested · Co-authored. Choose per task. Say why in one line. Put it on the sheet. Students should never have to guess where AI stands.

THE FOUR POSITIONS

Unassisted

No AI on this one

The skill is built by doing it, and the doing cannot be delegated. This is not a punishment setting — it names a task whose value is the doing.

The student may: work from their own head, notes, and sources; no generative tool in the loop.

The catch: only honest if the task genuinely can’t be done without building the skill. Otherwise it’s a ban dressed as a position — and students can tell.

Cleared

AI allowed, for a named use

AI is permitted for a specific, bounded job you name in advance — brainstorming, checking an outline, translating a term, formatting.

The student may: use it for the cleared purpose, and say where they did.

The catch: “cleared” is not “anything goes.” Name what it is cleared for, or it quietly becomes the whole assignment.

Contested

AI, configured to push back

The load-bearing one. The tool is set up to make thinking harder — it withholds, argues, seeds errors, refuses. Using it produces the thinking instead of presupposing it. This is the move that carries a student to the Deepen rung.

The student may: work against a tool built to resist them, and show the argument they won.

The catch: it fails silently. A helpful chatbot is not contested — if you haven’t configured it to resist, you’ve just Cleared it. (The teacher has always been the contester; AI is one more supplier of resistance.)

Co-authored

AI, and neither could have done it alone

“I couldn’t have written this without you, and you couldn’t have written it without me.” The student and the tool each bring something the other can’t.

The student may: build with the tool openly, and account for what each of them contributed.

The catch: could a machine produce this alone? then it’s output. Could the student alone? then the machine was decoration. Co-authored means neither could have.

HOW TO CHOOSE, PER ASSIGNMENT

Two questions settle most assignments:

Could a machine produce this alone?

If yes, you’re about to grade output. Change the task, or move it toward Contested so the thinking is what gets produced.

Could the student produce it alone?

If yes and the tool added nothing, the machine was decoration. Either drop it (Unassisted) or ask for something neither could do alone (Co-authored).

The clock tells you which. Time saved on what you are not assessing is the point of the tool. Time saved on what you are assessing is the warning light. Same clock, opposite meanings — the position you set is what tells you which one you’re looking at.

POSITIONS, RUNGS, OR THE AI ASSESSMENT SCALE?

Three vocabularies, one decision. Pick a grain and stay in it — don’t run all three at a student.

The four positions (this page) — what students see, named on each assignment.
The four rungs (Foundations · Enhance · Deepen · Innovate) — how your practice grows. Teacher-facing. See Four Foundations.
The AI Assessment Scale — the finer-grained alternative to the positions. Not a third system; just a smaller grain.

Write your course AI policy → · Judgment, not prompts →

Nobody is climbing toward permissiveness. Every position, done with depth, arrives at the same graduate: someone who can say what a tool is for — and when it isn’t.

Regis Jesuit High School · Educational Technology · A.M.D.G.

HelpDesk@regisjesuit.com · 303.269.8080 · LAST VERIFIED — AUGUST 2026