The transparency statement

Recommended for: Any course, on any assignment handed in outside class; a good first step if you have not addressed AI at all yet

Background for the instructor

Detection does not work and cannot be made to work, and treating version history as evidence costs more trust than it recovers. What is left is asking — and asking works better than it sounds, provided nothing punitive hangs on the answer.

Keep it to one assignment to begin with. A statement on every submission becomes a formality that nobody reads and everybody resents; one, discussed afterwards, tells you what your students are actually doing.

Preparation

Say in advance what the statement is for and what it is not for: it informs your teaching, it does not affect the grade. Then hold to that. Do the same yourself — label the material in your own course that you generated with AI, and say why.

Procedure

1. Add three questions to one existing assignment: what did you use AI for, at what point, and what did it give you?
2. Ask for one further sentence: looking back, was that a good use?
3. Read the statements as a set rather than one by one. The pattern across the class is the finding.
4. Report back to the group on what you saw, without naming anyone.
5. Use that as the opening for agreeing what is and is not acceptable in the rest of the course.

Discussion questions

• Was there anything you hesitated to write down, and what made you hesitate?
• Did writing the statement change what you did, or only how you described it?
• What would a fair rule for this course look like, now that we can all see what is actually happening?

Assessment opportunity

Do not grade the statement. Grade the reflection later in the course, once students have something to compare against.

Reference

• The Norton Guide to AI-Aware Teaching (Vee, Watkins & Bruff, W. W. Norton, 2026), chapters 7 and 14; recommended by Derek Bruff as the single smallest step for instructors who feel behind.