Before you write an AI policy, decide what you are protecting
The syllabus is almost finished.
The grading breakdown is there. The absence policy is clear. The late-work clause has been revised three times. Then the cursor reaches the line labeled “AI Policy,” and the work stops.
A few words appear. They disappear. The cursor keeps blinking.
That pause does not mean an educator is behind or unprepared. It points to a real gap between the pressure to publish an AI policy and the deeper thinking that policy requires.
Most syllabus decisions grow from years of classroom experience. Educators know what happens when late work goes unmanaged. They have seen the difference between offering grace and creating confusion. Even when the policy has never been formally articulated, the judgment beneath it has been developing for years.
An AI policy asks for a different kind of decision. The tools are still changing, and students are using them in complicated ways. A single assignment may include appropriate support, questionable shortcuts, and genuine learning.
Writing one paragraph requires an educator to take a position on learning, authorship, effort, and integrity. Polished language cannot carry that weight until the thinking beneath it is clear.
Why borrowed AI policies fall short
Educators have received plenty of sample language. Department templates and professional forums are full of policies that look finished.
Many of them still leave the most important questions unanswered.
A vague policy might say that AI tools must be disclosed and used appropriately. The word “appropriately” sounds reasonable until a student asks what it means for a particular assignment.
Appropriate for which task? Appropriate for what part of the work? Appropriate according to which learning goal?
A policy that cannot answer those questions is still a placeholder.
A punitive policy creates a different problem. It may list prohibited tools and attach academic-integrity consequences to their use. That approach can ignore the fact that AI assistance may support one part of an assignment while undermining another.
Blanket language also makes honest conversation harder. Students who want to understand the boundary may decide that asking a question carries too much risk.
An aspirational policy can sound thoughtful while remaining too abstract to guide anyone. It may affirm critical thinking, responsible use, and integrity without explaining what those values require when an assignment is due on Tuesday morning.
These weaknesses do not reflect a lack of care. Educators have been asked to write policies under time pressure, often without a clear framework. The profession is still working through the same questions.
The challenge belongs to the system. The decision still lands on the teacher’s desk.
The policy is the output
The AI policy is not the beginning of the process. It is what becomes possible after the educator has decided what the policy needs to protect.
The first step is to name the learning goal.
What is the student meant to develop through this assignment? Is the work designed to strengthen analysis, practice a skill, produce an original argument, or demonstrate synthesis?
A clear learning goal makes the next decisions more concrete. If the assignment exists so students can practice a skill, AI should not perform the practice for them. If the assignment asks students to synthesize information, the educator must decide which parts of that synthesis belong to the student.
Context matters. A reasonable boundary in a research-writing course may not fit a creative-writing workshop or a lab-based science class.
That difference is not inconsistency. It is professional judgment responding to the learning in front of it.
The three questions in the CALM AI Framework
The CALM AI Framework organizes classroom decisions around three questions:
What belongs to AI?
What belongs to the teacher? and
How does learning stay at the center of the classroom?
The first question identifies the role the technology may play. AI might help organize information, generate practice examples, or support a logistical part of the work. Its role should remain visible and connected to the assignment’s purpose.
The second question protects the work that requires human judgment. The teacher decides which parts of the learning process must remain with the student. That may include productive struggle, original thought, interpretation, or the ability to explain how a conclusion was reached.
The third question keeps the decision grounded in learning. It moves the conversation away from blanket permission and prohibition. The educator can ask whether a particular use supports the learning goal or quietly removes the thinking the assignment was designed to develop.
These questions do not produce the same policy for every course. They create a consistent way to reach a decision that fits the course.
That is the value of a framework.
What principled boundary language sounds like
Policy language changes when it grows from clear thinking.
It becomes specific without trying to predict every possible tool or situation. It explains the principle behind the boundary, and it gives students enough information to make responsible choices.
Consider two approaches.
One policy states that using AI constitutes academic dishonesty and refers students to the institution’s integrity policy. The sentence communicates a consequence, but it does not explain the learning the rule protects.
Another policy explains that the course is designed to develop a particular set of analytical skills. Students complete the core analysis independently, while limited AI support may be permitted during designated preparatory tasks. Any permitted use must be disclosed.
The second policy gives the educator language for a real conversation. It also gives the student a reason for the boundary.
Students may still ask questions. That is healthy.
A strong policy creates enough clarity for those questions to become part of the learning conversation.
Thirty focused minutes can change the policy
Working through the framework does not require a semester. A focused thirty-minute review can provide the clarity an educator needs before returning to the syllabus.
That time can be used to identify the learning goal, examine the intellectual work, and decide where human judgment must remain visible. The final policy then reflects a position the educator can explain and defend.
This matters when a student asks whether AI can help with brainstorming. It matters when a student uses a tool in an unexpected way. It matters when an academic-integrity concern requires more than a quick reference to a prohibition.
A policy built on principle can hold through those moments. A borrowed policy statement often cannot.
Practical support for writing the policy
The free CALM AI Toolkit includes four practical resources for educators. Two are especially useful during syllabus revision.
The “Sample Boundary Language for Syllabi” provides adaptable language for different approaches to classroom AI. The examples help educators see how a learning philosophy can become a clear boundary.
The “Four Questions Reference Card” supports the thinking underneath that language. It gives educators a repeatable process they can use when planning a course, reviewing an assignment, or responding to a new AI tool.
The goal is not to copy a finished policy and hope it fits. The goal is to understand the decision well enough to make the language honest.
AI policies will need to change. Tools will develop, assignments will evolve, and educators will learn from the conversations that happen in their classrooms. A reliable framework gives them a steady way to revisit those decisions without starting over every semester.
Good policy begins with a clear purpose
Thirty-five years in classrooms have taught me that students respond differently when an educator can explain the reason behind a boundary.
The policies that earn trust are grounded in a purpose students can understand. An educator should be able to look a student in the eye and say, “This is why the boundary matters, and this is what I am trying to protect for you.”
That kind of clarity cannot be borrowed. It grows from thoughtful professional judgment.
The blinking cursor is not evidence of failure. It may be evidence that the educator understands how much the policy needs to carry.
Please download the free CALM AI Toolkit for:
the Sample Boundary Language for Syllabi,
the Four Questions Reference Card,
the Assignment Triage Template, and
the 30-Day Integration Roadmap.
Policy writing becomes clearer when your thinking has somewhere grounded to begin.
AI Disclosure Statement
A note on how this gets made: I use AI to help with research and early drafts, the same kind of grunt work that used to eat my Sunday afternoons. The ideas are mine. The judgment calls are mine. I read every piece before it goes out, and if a line doesn't sound like something I'd actually say to you, it gets cut. That's the whole point of the CALM AI Framework: AI carries the volume. I keep the thinking.