On this page

Private, in-browser worksheet

Draft a course AI-use statement

Choose the boundaries you intend to explain, then edit the generated draft against your course and institution rules. Nothing you enter is sent to or saved by this site.

Which kinds of AI assistance may students use?

Select only uses you have decided to permit for this course or assignment.

This is a drafting aid, not an institutional policy or compliance check. Review the final wording, accessibility and alternative routes with the appropriate local guidance.

Start with the skill the assignment is meant to assess

An AI course policy is clearest when it explains the reason for its boundaries. Begin with a specific learning outcome: what should a student be able to explain, interpret, calculate, design or create without having a tool do that thinking for them? Cornell’s assignment-design guidance recommends checking that assignments align with course outcomes before deciding whether and how generative AI can support learning. Columbia’s teaching center likewise advises that syllabus statements fit the goals and needs of the particular course.

That framing helps distinguish tasks that sound similar but assess different work. Using a tool to generate practice questions is not the same as using it to answer an assessed exam. Asking for feedback on a draft is not the same as submitting generated prose. Explain which distinction matters for your assignment rather than relying on a broad label such as “AI allowed” or “AI prohibited.”

Supporting reading: [1] [2]

Separate the course-wide rule from each assignment’s rule

Scroll the table sideways to view all columns.

An assignment-level policy matrix to complete in plain language
Assignment purposePossible boundary to specifyWhat students need to know
Low-stakes practiceWhether a tutor or chatbot may explain a concept or generate practice questionsWhat they should attempt first and how to check the explanation
Source-based paperWhether AI may suggest search terms, summarize a source, edit prose or draft claimsWhich sources they must read and how to document assistance
Lab or data taskWhether AI may explain code, suggest a method or transform dataWhich calculations and interpretations are assessed as the student’s own
Timed or independent assessmentWhich tools, if any, are available during the assessmentThe exact approved access route and what to do if it fails

Put a short baseline in the syllabus, then make any assignment-specific exception visible where students receive that task. A course might allow practice and brainstorming generally while prohibiting generated analysis in a lab report. Another might permit a tool for a designated comparison task but not for the independent reflection that follows it. Tell students which instruction governs when the two appear to differ, and give them a clear way to ask before they submit.

Cornell describes policy markers as aids, not complete explanations, and notes that different assignments may have different learning goals. If you use a label, acronym or icon, pair it with plain-language details about permitted actions, limits, attribution and rationale. Do not expect students to infer whether “editing” includes rewriting a claim or whether “research” includes accepting a generated summary.

These are prompts for local decisions, not recommended permissions. Adapt each row to the assessment and the rules that govern it. Remove any example that does not fit your course.

Supporting reading: [1] [4]

Name the actions instead of saying “AI use”

Students need to know what a boundary means at the level of an action. Consider brainstorming, asking for explanations, practising, finding candidate sources, summarizing readings, translating, checking grammar, rewriting sentences, generating code, analyzing data, producing images and drafting assessed text separately. You do not need to write a long inventory if only a few actions matter; be explicit about the ones that change the learning task.

A useful statement pairs a permitted action with its condition. For example: “You may ask an approved tool to generate practice questions after you have reviewed the assigned notes. Check each answer against those notes; do not submit the generated answers.” This is an original illustrative template, not a rule from a particular institution. Replace its tool, sequence and conditions with the ones that actually apply.

Supporting reading: [2] [4]

Choose a policy pattern, then customize the wording

Scroll the table sideways to view all columns.

Original syllabus statement patterns to adapt, not copy without review
PatternDraft languageDecision still needed
RestrictedUnless an assignment explicitly says otherwise, complete assessed work without generative AI assistance. Ask before using a tool for an activity the instructions do not address.Does this cover the intended learning outcome, permitted accessibility tools and any institution-wide exceptions?
Bounded usesFor this course, use AI only for the actions listed in each assignment. You remain responsible for the reasoning and evidence the task assesses. Follow the stated record and verification steps.Which actions are permitted, on which assignments, and what evidence of process is appropriate?
Broad use with disclosureYou may use approved AI tools for this course unless an assignment sets a narrower limit. Describe material assistance as instructed, verify claims and sources, and submit work you can explain.Which tools, uses and disclosure format are approved, and which tasks require independent work?

The following original drafts show how scope can vary. They are not universal clauses, legal wording or endorsements. Check the institutional policy, assessment design, accessibility needs, available tools and local support process before adopting any language.

Avoid threat-only wording that does not tell students what to do. Explain the expected process and where to seek clarification. Columbia’s sample statements range from broad permission to prohibition, while also stressing course-specific adaptation and discussion with students.

Supporting reading: [1] [2]

Explain verification, disclosure and privacy separately

If a use is allowed, state who checks the result. A student may need to verify a quotation, citation, calculation, translation or factual claim against an original source. Cornell’s guidance tells instructors to clarify documentation and attribution expectations when AI is permitted and reminds students to verify output and references. A disclosure records how a tool was used; it does not prove that the output is accurate or that the use was allowed.

Give concrete documentation instructions when a record is required: what details to include, where to put them, and whether prompts, retained output or an appendix are needed. Avoid saying “cite AI” when the intended requirement is a process note, or suggesting that one citation format satisfies every course policy. Our separate guide explains how citation and disclosure differ.

Set a data boundary before students upload work or course materials. Name approved accounts or tools where relevant, identify material that must not be entered, and point to the institution’s privacy or IT guidance. UNESCO’s education guidance places privacy, age-appropriate use and human agency among the issues that policy and pedagogical design should address. Its general guidance does not replace local rules or a service’s current terms.

Supporting reading: [1] [3] [5]

Communicate the rule before the assignment is due

Publish the course baseline in the syllabus or course site, repeat the relevant rule in each assignment prompt, and explain how students can ask a question. Use consistent terms across the syllabus, rubric and class discussion. If an assignment changes the default, make that exception easy to spot. Ask students to describe an uncertain use before they rely on it, and update the wording when the task or approved access route changes.

A short conversation can expose ambiguous verbs such as “assist,” “edit” or “support.” Give an example of an allowed request and a nearby request that crosses the boundary, then ask students to explain the difference. If a student needs an alternative or accommodation, follow the institution’s established process rather than making the policy builder decide what is appropriate.

Supporting reading: [1] [2]

Copy-ready course AI policy template

Use this outline if you want to write the statement yourself. Replace every bracketed phrase and delete sections that do not apply. It is an original editing scaffold, not a ready-made institutional policy.

“In [course or assignment], you may use [named tools or actions] for [specific purpose]. You may not use AI to [specific prohibited action]. You are responsible for [reasoning, evidence, calculations or other assessed work]. When you use a permitted tool, [state the required record or disclosure and where to submit it]. Verify [claims, quotations, references or outputs] against [the assigned material or original sources]. Do not upload [sensitive, confidential, unpublished or restricted material] unless the approved institutional route permits it. For an unclear case, contact [person or channel] before submitting. Assignment instructions may give a more specific rule; [explain how students can identify the controlling instruction].”

Supporting reading: [1] [2] [4]

Questions about this topic

What should an AI policy in a syllabus include?

State the course-wide default, name permitted and prohibited actions, explain assignment-specific exceptions, describe any disclosure or verification requirement, set a privacy boundary and tell students how to ask about an unclear case. Adapt the statement to the learning outcomes and institutional rules.

Should AI be allowed for every assignment in a course?

There is no universal answer. Decide based on the skill each assignment assesses, the practice students need and the rules governing the course. Different assignments can have different boundaries if the instructions make them clear.

How do I explain AI use in an assignment prompt?

Name the action rather than using a broad label. Say whether students may brainstorm, practise, locate candidate sources, edit language, generate material or use another specific function. Then explain any conditions, checks and documentation.

Do students need to cite AI whenever they use it?

Citation and disclosure are different. Requirements vary by course, style guide and type of use. Tell students which record you expect and where to place it; do not assume an AI citation alone documents the process or grants permission.

Can I use the generated policy as my official syllabus rule?

Not without review. It is a draft assembled in the browser, not a legal or institutional policy check. Verify it against your institution, course outcomes, assignment instructions, approved tools, privacy requirements and accessibility process.

Does the policy builder save the information I enter?

No. The builder creates text in the current browser tab and offers local copy and download actions. It does not send the form to this website or save it automatically.

Sources and editorial notes

The supporting reading below informs our editorial guidance. Exercises and scenarios illustrate a method; they do not report a completed product test or measured learning outcome. Read our editorial policy or send a correction with a supporting source to info@ai-essay.com.

  1. Columbia Center for Teaching and Learning: Developing AI Syllabus Statements and Communicating with Students about AI
  2. Cornell Center for Teaching Innovation: AI in Assignment Design
  3. Cornell Center for Teaching Innovation: AI & Academic Integrity
  4. Cornell Center for Teaching Innovation: AI Course Policy Icons
  5. UNESCO: Guidance for generative AI in education and research