A study of roughly 31,000 syllabi from one large public research university found that instructors have been moving away from broad AI restrictions and toward distinguishing which specific activities benefit from AI and which require independent work. Courses that rely heavily on writing and coding were significantly more likely to include an AI policy, which suggests that instructors respond strategically to the tasks their students perform. The finding matches what many writing teachers experience: a blanket rule rarely fits the full range of assignments in a course.

A rule that says AI is banned leaves students guessing about brainstorming, grammar checking, or asking a chatbot to explain a confusing reading. A rule that says AI is allowed leaves instructors without protection when a student submits a paper they did not write. Both versions fail for the same reason, which is that they describe a tool instead of describing the learning each task is meant to produce.

Task-based policies solve that problem by attaching the rule to the assignment. The syllabus names the skill being assessed and then says what kinds of help support that skill and what kinds replace it. Students can then make honest decisions, and instructors have a clear standard to point to if a conversation about a particular submission becomes necessary. A policy permitting AI for brainstorming but not for drafting a thesis is far easier to follow than a blanket ban that nobody believes.

How to write the policy

Start by listing the major assignments in your course and asking one question about each: what must the student be able to do without help by the end. For an argument essay, the answer might be forming a claim and organizing evidence, which means AI cannot generate the thesis or body paragraphs. For a research proposal, you might allow AI to suggest search terms but require students to read and summarize every source themselves.

• Name the core skill each assignment is designed to measure in one plain sentence.
• List permitted uses, such as explaining a concept, quizzing the student, or checking grammar after drafting.
• List prohibited uses, such as generating thesis statements, paragraphs, or reflections for submission.
• Require a one-sentence disclosure of any AI use at the end of each assignment.
• State what happens if the policy is misunderstood, such as a revision opportunity before any penalty.

Discipline differences are real

The same research found clear differences by field, with business courses among the most permissive and humanities courses among the most restrictive. Those differences are not arbitrary, because a humanities seminar is often built around the act of forming an argument in sentences, while a business course may treat AI-assisted analysis as a professional skill. A writing program that spans several departments should expect and respect those differences rather than forcing a single policy on every course.

Where departments share students, such as a first-year writing program feeding many majors, it helps to publish a common vocabulary even if the rules differ. Terms like brainstorming, drafting, revising, and editing can be defined once and then used consistently, so that students moving between courses are not confused by different meanings. A short shared glossary costs little and prevents many misunderstandings.

Keeping grading fair under a task-based policy

A task-based policy works best when grading reflects the skill being assessed. If the rubric rewards claim quality, evidence selection, and reasoning, then a polished but shallow paragraph does not earn a high score regardless of how it was produced. Rubric-aligned feedback, whether written by hand or drafted by a tool and reviewed by the instructor, tells students exactly where their thinking needs to improve.

Pair the policy with process evidence, such as a brief planning note or a draft submitted a week early, so that authorship questions can be answered with documents instead of suspicion. Students who know the process is part of the assignment tend to engage in it honestly. Instructors, in turn, spend less time investigating and more time teaching. A planning note or an outline with timestamps takes students only a few minutes to produce and gives you a reliable record to consult.

Review and revise each term

Treat your policy as a living document and revisit it before every term. Ask students which rules were clear, which seemed unreasonable, and which uses they were unsure about, and adjust the language accordingly. Tools and student habits change quickly enough that a policy written two years ago may be answering questions no one is asking anymore. Even a ten-minute conversation with a handful of students at midterm can reveal confusion that a revised paragraph in the syllabus would resolve.

Share your policy with colleagues, especially those who teach the same students, so that expectations feel coherent across courses. Departments that compare policies often discover gaps and contradictions they can resolve together. Over time, a shared, task-based approach builds a culture in which students understand why the rules exist and what they are protecting. Students notice when one instructor allows grammar checking while another forbids it, and a short shared glossary of terms can prevent that confusion.