Last updated: October 2026 · Reading time: about 12 minutes
Many classrooms now face the same problem: students have access to powerful AI tools, but the rules for using them are unclear. Unclear rules create unfair accusations, inconsistent grading, and stress for everyone. This guide gives you a practical AI policy template for teachers and schools, with sample wording you can copy and adapt, a model for levels of allowed AI use, rules for how to treat AI detector results, a student disclosure form, and a rollout checklist.
Table of contents
- Why every course needs an AI policy
- Five levels of AI use
- Sample policy wording (copy and adapt)
- How to treat AI detector results
- A fair investigation process
- Student AI disclosure form
- Rollout checklist
- FAQ
Why every course needs an AI policy
- Clarity. Students cannot follow a rule that does not exist. Silence on AI makes accusations harder to defend.
- Fairness. Without written criteria, two instructors can treat the same behavior differently.
- Learning. Some AI use can support learning when it is guided and transparent. A policy lets you decide where it helps and where it harms the skill being assessed.
- Due process. If detector scores are used, a policy should explain their role and limits. See our guide to AI detectors as evidence.
Always check whether your school or district already has an official policy. Your course policy should fit inside it, not contradict it.
Five levels of AI use
A simple scale lets you set a level for each assignment, so students know exactly what is allowed. Adapt the labels to your context.
| Level | Name | What is allowed |
|---|---|---|
| 0 | No AI | All work must be produced without generative AI. Used for in-class writing, exams, and skills tests. |
| 1 | Planning only | AI may help brainstorm topics or questions. No AI-generated sentences in the submission. |
| 2 | Feedback only | AI may give feedback on a student's own draft (clarity, structure). Student makes all changes. |
| 3 | Limited drafting | AI may produce limited text for specified sections. Student must edit, verify, and disclose. |
| 4 | Open use with disclosure | AI use is encouraged within the task. Student must document prompts and verify all claims. |
Sample policy wording (copy and adapt)
The text below is a starting point. Replace bracketed items and adjust to your institution's rules.
1. Purpose. This policy explains how generative AI tools (such as chatbots and writing assistants) may be used in [Course Name]. Its goal is to support learning, protect academic integrity, and treat all students fairly.
2. Assignment levels. Each assignment states an AI-use level from 0 to 4. If no level is stated, Level [1] applies. If you are unsure, ask before submitting.
3. Disclosure. If you use AI in any way permitted by the assignment, you must complete the AI Disclosure Statement, describing the tool, the purpose, and how you changed or verified the output.
4. Responsibility. You are responsible for the accuracy and originality of everything you submit. AI tools can produce false information and invented sources. You must verify all facts and citations.
5. Prohibited conduct. Submitting AI-generated work as your own where it is not permitted, fabricating sources, or using tools to disguise AI-generated text violates the academic integrity policy.
6. Evidence of process. You may be asked to provide drafts, notes, version history, or to discuss your work. Keep these materials until the end of the term.
7. Detection software. AI detection tools can produce errors. A detector score will not be used as the sole basis for any finding of misconduct. See Section 8.
8. Process. If a concern arises, the instructor will explain it in writing, give you the chance to respond and provide evidence, and consider all information before any decision. You may request review under [Institution's Appeals Procedure].
How to treat AI detector results
Detectors estimate whether text resembles AI output. They cannot prove authorship, and research has documented false positives, including higher rates for some groups of writers such as non-native English speakers (see our article on AI detectors and non-native English speakers). A responsible approach looks like this:
- State the role of detectors in writing. For example: "Detector results may prompt a conversation but are not evidence of misconduct by themselves."
- Use more than one source of evidence. Drafts, version history, comparison with earlier work, and an oral discussion carry far more weight.
- Be transparent. If you use a detector, tell students which tool and how results will be used.
- Protect privacy. Do not upload student work to a tool unless your institution has approved it and its data practices are clear.
- Train instructors. Make sure staff understand false positives and the limits of scores.
A fair investigation process
- Private, non-accusatory conversation. Describe what you noticed and ask the student about their process.
- Request process evidence. Drafts, notes, sources, and version history.
- Compare with prior work. Consider differences carefully, since writing changes as students improve.
- Offer an oral or supervised check. A short discussion of the thesis and sources can be informative and fair.
- Document everything. Keep notes of what was asked, answered, and decided.
- Decide based on the total evidence. If it is inconclusive, treat it as inconclusive.
- Explain the right to appeal. Provide the procedure in writing.
Student AI disclosure form (sample)
Assignment: ______________________ Date: ____________
1. Did you use any AI tool for this assignment? ☐ No ☐ Yes
2. If yes, which tool(s)? ______________________________
3. What did you use it for? (brainstorming, outline, feedback on my draft, grammar, other) ______________________________
4. How did you check, edit, or verify the output? ______________________________
5. I confirm that this statement is accurate and that the submitted work reflects my own understanding. Signature: ______________
Rollout checklist
- ☐ Confirm institutional rules and align the course policy.
- ☐ Choose an AI-use level for every assignment.
- ☐ Put the policy in the syllabus and the learning platform.
- ☐ Spend one class session discussing examples of acceptable and unacceptable use.
- ☐ Provide the disclosure form with each assignment.
- ☐ Add low-stakes process checkpoints (outline, draft, reflection).
- ☐ Decide, in writing, how detector results will and will not be used.
- ☐ Review the policy at the end of the term and update it.
Watch: AI policy ideas from educators
Frequently asked questions
Do I need an AI policy if my school has none?
Yes. A clear course-level policy protects students and teachers, even if you later need to align it with a school-wide rule.
Should I ban AI completely?
That depends on the learning goals. For skills such as in-class writing, a ban may be appropriate. For other tasks, structured use with disclosure can be more realistic and educational. Many teachers use different levels for different assignments.
Can I fail a student because of an AI detector score?
Treating a score as sole evidence is risky and often against institutional policy. Combine it with other evidence and follow a fair process.
How can I reduce AI misuse without detectors?
Use process-based assessment: outlines, drafts, in-class writing, personal or local topics, oral defenses, and reflective notes.
Is this template legal advice?
No. It is general sample wording. Review it against your institution's rules and legal requirements before use.
Key takeaways
- Set one AI-use level per assignment and state it clearly.
- Require disclosure and process evidence instead of relying on detection alone.
- Write down how detector results will and will not be used, and guarantee a fair response process.
Sources and further reading
- Dr. Sarah Elaine Eaton, "The Use of AI-Detection Tools in the Assessment of Student Work".
- Liang et al., "GPT detectors are biased against non-native English writers".
- Bloomberg, "AI Detectors Falsely Accuse Students of Cheating—With Big Consequences".
Disclaimer: This template is provided for general information and adaptation. It is not legal advice. Adapt it to your institution's policies and applicable laws.