Three-Tier AI Policy vs. the CALM AI Framework: What Districts Still Need
A three-tier language has been spreading through school districts’ AI policy over the past year: tasks are labeled AI-prohibited, AI-assisted, or AI-collaborative, and the label tells a teacher, a student, and a parent exactly where a given assignment sits. More than 200 districts have adopted some version of this model. It is easy to see why. A three-tier label fits on one page; it fits in a board presentation; and it gives a nervous staff something concrete to point to on a hard day.
The three-tier language is also solving a real and urgent gap. Recent industry research on AI adoption in K-12 found that a majority of high school students have already used AI on a graded assignment, while only about a third of schools have any formal AI policy at all. That gap between use and guidance is exactly where confusion, inconsistency, and mistrust grow. A three-tier system is a reasonable, fast answer to "we need something in writing before this gets worse."
The trouble shows up the moment a real assignment lands on a real Tuesday.
What a Tier Label Cannot Do
A tier is a category. It answers what is allowed for a type of task and was decided only once (usually at the district or department level) and often before the specific assignment exists. That works cleanly for the easy cases: a state exam is prohibited, a brainstorming tool is assisted, a co-designed project is collaborative. Unfortunately, most classroom decisions are not the easy cases.
The same assignment can sit in a different tier depending on the learning goal behind it. A teacher building fluency with thesis statements needs that skill built without help, even if the district's general policy calls essay-writing tools "assisted" for that grade level. A teacher whose goal is idea generation, not sentence-level writing, might reasonably let AI further into the same-looking task. The tier was decided in a policy meeting. The learning goal is decided by the teacher, in their unique classroom, and assignment by assignment. A label written in advance cannot see that far ahead, and it was never built to.
This is not a flaw unique to the three-tier model. It is what happens to any fixed category system applied to a moving target. New tools appear faster than any tier list can be updated, and a policy document written last spring is already behind the tools students will be using this fall.
What Works In the Classroom Instead
The CALM AI Framework was built to answer a different question. Instead of sorting tasks into categories in advance, it gives a teacher three questions to carry into any classroom decision, regardless of which tool showed up or which tier it was assigned last year:
What is the learning goal? Named precisely, not in general terms.
What belongs to AI, and what belongs to the student? Decided against that specific goal, not against a generic tool description.
How does this decision keep learning at the center? Checked against the goal one more time before the tool comes out.
These three questions do not replace a tier system. They make one usable. A district can keep its three-tier language exactly as written, useful for parent letters, board minutes, and staff training shorthand, while giving teachers a way to decide, in the actual moment, which tier a specific task belongs in for a specific student and a specific goal. The tier answers what the district allows in general, and the three questions answer what this assignment, this week, actually needs. Districts do not have to choose between the two: one communicates, the other decides.
Where this Fits in a District's Implementation
A one-page policy, however well built, is a starting document not a finished one. The CALM AI Sprint takes a district's existing AI policy (tier system included where one already exists) and builds the living framework, decision practice, and staff-wide judgment underneath it. The CALM AI Spring includes three live sessions, a custom-configured decision bot staff can use 24/7, up to 30 redesigned assignments with rationale, and a full staff readiness survey, all built around the same three-question protocol described here.
The CALM AI Framework itself is available as a free one-page download, built for any individual teacher who wants the three questions on a desk before Tuesday morning's decision arrives. It works alongside a district's existing policy language (three-tier or otherwise) rather than asking anyone to throw that language out.
Download the CALM AI Framework here, and for districts building or refining a full AI policy, more on the CALM AI Sprint is here.
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.