Writing Your Course AI Policy

Novice · Teaching & Learning track · ~35 min hands-on + readings and quiz

← Back to the Teaching & Learning track

A clear, enforceable AI policy for your syllabus — not a disclaimer, a working document.

What you’ll be able to do

  • Write a course AI policy that is specific, enforceable, and fair
  • Adapt the policy to different assignment types within the same course
  • Anticipate student questions and edge cases before the semester starts

Overview

A course AI policy tells students what’s allowed, what isn’t, and what happens when the line is crossed. Most existing policies are either too vague (“use AI responsibly”) or too broad (“AI is prohibited in this course”). Neither works well — vague policies are unenforceable, and blanket bans are unrealistic and may disadvantage students who need accessibility tools.

A good policy is assignment-specific, states what must be disclosed, and explains why — connecting the rules to the learning objectives. Students follow policies they understand.

You’re in good company: funders (NIH, NSF) and journals (ICMJE, COPE) have recently written their own specific, enforceable AI-use rules for grant applications and manuscripts. Pointing students to these is a useful move — they’re real-world examples of what “specific and enforceable” looks like in the professional settings students are headed for.

Practice activities

Activity 1 · Novice — Draft your course AI policy

Time ~18 min · Tools ChatGPT Edu

Goal. Produce a syllabus-ready AI policy tailored to your course.

Setup. Use this sample course, or substitute your own:

Course: EPID 602 — Intermediate Epidemiology (MPH required course). Assignments: weekly problem sets (calculations and interpretations), a midterm case study analysis, a final research proposal, and in-class participation. Learning objectives: Apply epidemiologic measures, interpret study designs, and critically evaluate published research.

Steps.

  1. Draft a policy:

    Draft a course AI policy for this epidemiology course. The policy should: (a) state the overall stance (AI is a tool, not a replacement for learning), (b) specify allowed and prohibited uses for each assignment type, (c) require disclosure of AI use, (d) explain consequences for violations, and (e) include 2-3 specific examples so students know where the line is. Keep it under 400 words — students won’t read a 2-page policy.

  2. Review for enforceability: Could you actually detect a violation? If you can’t tell whether a student used AI for their problem set calculations, your prohibition may be unenforceable — consider redesigning the assignment instead.

  3. Add a “why” statement — one sentence explaining how the policy connects to the learning objectives:

    Add a one-sentence explanation of why this policy exists, framed in terms of the course learning objectives — not “because it’s cheating” but because of what students need to be able to do.

Expected result. A syllabus-ready policy under 400 words that distinguishes among assignment types and includes specific examples.

Check your work. Have a colleague or TA read it. If they can identify a scenario where the right answer isn’t clear, tighten the language.

Common pitfalls. The most common mistake is writing one rule for the whole course. Problem sets, case studies, and research proposals have different learning objectives — the AI rules should differ too.

Stretch (optional). Draft a short FAQ (3–4 questions) anticipating the student questions you’ll get in the first week.

Activity 2 · Novice — Stress-test and compare policies

Time ~15 min · Tools ChatGPT Edu

Goal. Find the ambiguities in your policy before your students run into them.

Setup. Your draft policy from Activity 1.

Steps.

  1. Stress-test it:

    Read this course AI policy as a student acting in good faith who genuinely can’t tell what’s allowed. List every place where a reasonable, honest student could read the policy two different ways, and say what question they’d have to ask me to resolve each one.

  2. Close the gaps — revise the language so each of those questions is answered in the policy itself.

  3. Compare to a different approach:

    Now draft an alternative policy for the same course that takes the opposite stance — one that actively integrates AI into assignments and assesses students on how well they use it. What changes?

  4. Decide which approach (restrictive, integrative, or a hybrid) fits your learning objectives better.

Expected result. A tightened policy plus an alternative version that shows the other end of the spectrum. Most courses land somewhere in between. (For the same exercise from the other side of the desk — students reading three different policies in one term — see AI in Your Coursework, the student-facing mirror of this page.)

Check your work. The final policy should pass three tests: (1) A student can read it and know what to do. (2) You can detect most violations or have designed assignments where violations don’t bypass learning. (3) It connects to your learning objectives, not just institutional rules.

Common pitfalls. Banning AI use in a course where students will use AI in their jobs after graduation deserves a justification. “Because it’s academic dishonesty” is weaker than “because you need to be able to do this calculation yourself in a job interview.”

Stretch (optional). Identify one assignment where integrating AI use is the learning objective — e.g., “Use AI to draft a methods section, then annotate where it got your study design wrong.”

Check your readiness

Answer these, then check — your score suggests whether to dive in or skim the readings first.

Useful resources

Give Feedback on This Page

NoteTwo minutes: tell us how this track is working

Your feedback shapes which topics get expanded and what new activities are added.

Self-paced track feedback form ↗

In the form, select this page from the list of topics.