The Syllabus Clause Wars

Take a typical first-year student with five courses. Odds are good they’re operating under several different AI policies at once. In one class, using a chatbot for anything is misconduct. In another, it’s encouraged for brainstorming but banned for drafting. In a third, the syllabus doesn’t mention it at all.

None of these instructors is wrong to set their own rules. But the student is the only person who has to hold all five in their head — and the only one who pays when they get it wrong.

The three families of policy

Most course AI policies fall into one of three broad shapes.

The ban. Any use of generative AI is treated as an integrity violation. Simple to state, very hard to enforce, and increasingly difficult to define as AI features appear inside word processors, search engines and grammar tools.

The open door. AI is allowed, sometimes expected, with a requirement to disclose how it was used. This shifts the question from “did you use it” to “did you use it honestly.”

The traffic light. Each assignment is labeled: red for no AI, amber for limited uses like brainstorming or proofreading, green for anything goes with disclosure. Variations of this model have spread widely because they let instructors match the rule to the learning goal.

Where it breaks down

The traffic-light approach is the most thoughtful of the three, but it has its own failure mode: ambiguity at the edges. Is a grammar checker that rewrites a sentence “proofreading” or “drafting”? Is asking a chatbot to explain a reading “research” or “outsourcing”?

Ask students and many will tell you they’d follow the rules if they understood them. Clear rules are often what’s missing.

Detectors make it worse

Add AI detection to an inconsistent policy landscape and the confusion compounds. A detector doesn’t know which assignment was amber and which was red. It doesn’t know a student was permitted to use AI for an outline. It produces a score, and the score lands in a context it can’t see.

What good policy looks like

The clearest policies we’ve seen share a few traits:

1. They’re assignment-level, not course-level. The rule sits right next to the task.
2. They give examples of allowed and forbidden uses, not just categories.
3. They define disclosure — what to say, and where to say it.
4. They explain the why. “This essay is practice for the exam, where you won’t have AI” is a reason students can respect.

Clear rules won’t end the debate over AI in education. But they would end a lot of the accidental cases, where students who were trying to do the right thing simply guessed wrong about which class they were in.