Your department’s AI policy has to say something specific about three different kinds of AI, in every course component. This session does one component, by hand, in about fifteen minutes. You finish holding a paragraph you can paste into a syllabus today.
Nothing here is generated, saved or sent anywhere. This page runs entirely in your browser. Copy your paragraph out before you close the tab, because when you close it, it is gone.
Seed. Name the one component
Not the course. One graded thing inside it. The essay, the lab report, the discussion post, the problem set, the clinical reflection. Pick the one you are most unsure about.
One sentence. Every rule below is downstream of this answer, so write it first.
Sprout. The generative rule
Generative AI produces the artifact. A draft, a paragraph, an image, a block of code. This is the column your current syllabus almost certainly addresses. Write the rule as a permission or a boundary, not as a warning.
Grow. The two columns nobody has written yet
Assistive AI supports the person doing the work. Spelling and grammar correction, read aloud, live captioning, translation, predictive text, reading level adjustment. Before you write this rule, look at your own accessibility statement. For some students in your room these are the accommodation, not a shortcut.
Agentic AI takes actions. It opens the sources, runs the code, fills the sheet, drafts and sends, and reports back on what it did. The student sets a goal and may never touch a step. This is the hard one. The useful question is not whether it is allowed. It is what the student has to be able to show you afterward.
Harvest. One line, and one decision
A policy a student cannot act on is a policy you will end up arguing about in week nine. Write the single sentence a student must include with the submission. Keep it short enough that they will actually write it.