"AI policy offers the starting line. The teachers who shape how students use AI — with purpose, honesty, and real thinking — are the ones who build students capable of using it well."
This toolkit goes beyond tiers and policy — giving teachers a complete, practical workflow for setting clear AI expectations, building assignment-specific language, and making AI use visible, accountable, and grounded in real learning. All in one place, ready to copy.
Select a grade band, choose your AI tier, pick the specific AI actions students may use, and set your citation requirements. The builder generates your assignment language, rubric, and student prompt template — ready to copy.
Grade band shapes the documentation requirements, rubric language, and assignment language tone so the outputs actually fit your students.
Expand any section below for detailed frameworks, examples, and response strategies.
Syllabus language establishes the class-wide norm students carry into every task. Two versions below — pick the one that fits.
Each tier defines how much AI a student may use on a given assignment. Select a tier, then choose specific approved actions from the task menu. If no tier is stated on an assignment, students should assume Tier 1.
These are the same tiers and actions used in the Assignment Builder tool above.
Tier framework adapted from the AI Assessment Scale (Perkins et al., 2024) under CC BY NC SA 4.0.
Before writing the assignment, consider where AI can shortcut the learning — and how to close those gaps at the design stage. Run your existing prompt through an AI tool first. If it can get a passing grade, revise before it reaches students.
Homework is where AI shortcuts happen most. The fix isn't more control — it's shifting focus from final product to preparation.
Students include a brief one-sentence Process Note with their assignment submission to document their AI use. This acknowledgment is designed to limit documentation fatigue while practicing responsible AI use.
Primarily used for everyday assignments where AI played a limited, mechanical role — editing, grammar, or outlining — and the student's own writing represents the final work.
Students cite AI as they would any source — because it is a source when it contributes ideas, structure, or language to their work. This teaches academic integrity aligned with higher education's emerging standards.
Students must include two things: a formal citation in their required style, and a brief attribution statement explaining what AI contributed and what the student revised, verified, or wrote themselves.
Students maintain a complete record of all AI prompts and outputs throughout their assignment process. Documentation is submitted alongside the final work — in a Google Doc or teacher's preferred format.
This level asks students to do more than note AI use — it asks them to evaluate it: what AI got right, what it missed, what they corrected, and how they ensured the final work reflects their own thinking.
Thinking-based assessments bring learning back into view. They allow teachers to observe reasoning, uncertainty, and adaptability in ways that polished final products never can.
| Learning Goal | Best-Fit Assessment Strategies |
|---|---|
| Conceptual Understanding | Cold Start Checks, Error Analysis, Transfer Problems, Whiteboard Reasoning |
| Skill Application | Transfer Problems, Field Research, In-Class Writing, Analog Assignments |
| Metacognition & Self-Awareness | Oral Assessment, Process-Weighted Rubrics, Error Analysis, In-Class Writing |
| Communication & Discourse | Socratic Seminar, Structured Debate / Mock Trial, Oral Assessment |
| Real-World Inquiry | Field Research, Transfer Problems, Analog Assignments |
| Independent Reasoning | In-Class "Invisible Work," In-Class Writing, Cold Start Checks, Whiteboard Reasoning |
Sources: MIT Media Lab (Kosmyna et al., 2025) · Carnegie Mellon & Microsoft Research (Lee et al., 2025) · BJEST (Fan et al., 2025) · Lang, J.M. Small Teaching, 2021 · Bjork & Bjork, 2011 · University of Sydney (Liu et al., 2025)
Compare what you're seeing against the student's established work. Look for patterns across multiple indicators.
When a student admits to misusing AI, the response should be restorative, proportional, and focused on learning. The goal is to help students understand why the shortcut harmed their learning — not just that they broke a rule. Consider collaborating with colleagues.
Not every case is straightforward. Students may have used AI in ways they didn't realize were misuse, partially relied on it while adding their own thinking, or used AI for an early step and then completed the writing themselves.
Documentation protects both the student and the teacher. Keep brief, factual records:
Informed by: RAND Corporation (Doss et al., 2025) · PACE/TeachAI Policy Workgroup (2023–2025) · International Center for Academic Integrity · MIT Teaching Systems Lab (Smith et al., 2025) · ISTE Standards for Educators · eSchool News (2025)
A 10-part prompt framework and 16+ ready-to-use templates covering everyday teaching tasks (lesson plans, rubrics, assessments, grading, parent communications, sub plans) and AI-aware assignment design (shortcut audits, redesign for thinking, Bloom's alignment). Quick-use and extended versions included.
Open the Educator Prompt Toolkit →The template in your outputs gives students structure. But AI doesn't know what tier you chose — and without a constrained prompt, it will give more than you approved. The Student Prompt Builder walks students through building their own prompt with your limits baked in.
This is optional — not every assignment needs it. But for assignments where AI use is significant (Tiers 3–5), it's the best way to protect effortful learning while still letting students use the tools.
Open the Student Prompt Builder →