Policy builder: pick a level of AI use
Choose how much AI use an assignment allows. Each level comes with a draft syllabus statement and published examples from university teaching centers.
- Pick a level for each assignment; levels describe conditions of use, not quality.
- The builder writes a student-facing statement and private instructor notes.
- Check your institution’s policy before you publish a statement.
What you get
Pick a level and answer a few questions. The builder writes a student-facing statement you can paste into an assignment, plus private notes for your planning.
Go to the builder ↓For this essay, you may use AI to brainstorm topics and find sources to check. You must write the argument and all drafts yourself. AI may not draft or rewrite your text. At the end, add two or three sentences saying which tool you used and how. Verify every source AI suggests against the original. Do not paste classmates’ work or personal information into AI tools. If you are unsure whether a use is allowed, ask me first.
About the levels. Level names follow the AI Assessment Scale (Perkins, Roe & Furze). The levels describe conditions of AI use, not rising educational quality: a higher level is not a better assignment. Level 4 allows AI throughout a task with a set goal. Level 5 is for tasks designed around exploring what AI can do.
No AI
AI may not be used at any stage. Students complete the work entirely on their own, often in supervised or in-class settings.
Works well for: foundational skills, in-class writing, exams and tasks where independent ability is what you assess.
Draft syllabus statement
You must complete all work for this assignment yourself, without the use of generative AI tools at any stage, including brainstorming, drafting and editing. This helps you build the foundational skills the course is designed to develop. If you are unsure whether a tool counts as AI, ask me before you use it.
Adapted from the published examples below and the AI Assessment Scale.
Published statements for Level 1
Short excerpts from university teaching centers. Follow the links for the full statements.
“You should complete all graded assignments in this course yourself, without any use of generative artificial intelligence (AI).”
“Work submitted for a grade by students, including all drafts, brainstorming, and final projects, will be created without the use of generative AI.”
“The use of AI tools… is prohibited for any part of the assignment process, including brainstorming, writing, and editing.”
Mixing levels in one course
Many published templates combine a default with exceptions set per assignment. Carnegie Mellon’s example states that permitted AI use will be named in the assignment prompt, and otherwise the default is that generative AI is disallowed. Recent syllabus research finds instructors moving the same way, from course-wide bans to task-based rules (Chirikov, 2026).
What you get
Pick a level and answer a few questions. The builder writes a student-facing statement you can paste into an assignment, plus private notes for your planning.
Go to the builder ↓For this essay, you may use AI to brainstorm topics and find sources to check. You must write the argument and all drafts yourself. AI may not draft or rewrite your text. At the end, add two or three sentences saying which tool you used and how. Verify every source AI suggests against the original. Do not paste classmates’ work or personal information into AI tools. If you are unsure whether a use is allowed, ask me first.
About the levels. Level names follow the AI Assessment Scale (Perkins, Roe & Furze). The levels describe conditions of AI use, not rising educational quality: a higher level is not a better assignment. Level 4 allows AI throughout a task with a set goal. Level 5 is for tasks designed around exploring what AI can do.
AI Planning
AI may support early thinking, such as brainstorming, outlining or finding sources, but the drafted work must be the student’s own.
Works well for: research projects and essays where idea generation is not the main learning goal.
Draft syllabus statement
You may use generative AI to brainstorm, plan or explore sources for this assignment. All drafted writing and analysis you submit must be your own. Include a short note describing how you used AI in planning, and keep any AI conversations in case I ask to see them.
Adapted from the published examples below and the AI Assessment Scale.
Published statements for Level 2
Short excerpts from university teaching centers. Follow the links for the full statements.
“You can choose to use AI tools to help brainstorm assignments or projects or to revise existing work you have written.”
“Not use AI tools to replace your own thinking or analysis or to avoid engaging with the course content.”
“Not permitted… Asking AI to outline or structure your assignment before you have developed your own approach.”
Mixing levels in one course
Many published templates combine a default with exceptions set per assignment. Carnegie Mellon’s example states that permitted AI use will be named in the assignment prompt, and otherwise the default is that generative AI is disallowed. Recent syllabus research finds instructors moving the same way, from course-wide bans to task-based rules (Chirikov, 2026).
What you get
Pick a level and answer a few questions. The builder writes a student-facing statement you can paste into an assignment, plus private notes for your planning.
Go to the builder ↓For this essay, you may use AI to brainstorm topics and find sources to check. You must write the argument and all drafts yourself. AI may not draft or rewrite your text. At the end, add two or three sentences saying which tool you used and how. Verify every source AI suggests against the original. Do not paste classmates’ work or personal information into AI tools. If you are unsure whether a use is allowed, ask me first.
About the levels. Level names follow the AI Assessment Scale (Perkins, Roe & Furze). The levels describe conditions of AI use, not rising educational quality: a higher level is not a better assignment. Level 4 allows AI throughout a task with a set goal. Level 5 is for tasks designed around exploring what AI can do.
AI Collaboration
AI may help draft and refine work, as long as students critically evaluate its output, revise it themselves and disclose their use.
Works well for: writing-intensive and project courses that teach students to work critically with AI.
Draft syllabus statement
You may use generative AI to help draft and revise this assignment. You are responsible for checking every claim, source and calculation, and the final work must reflect your own judgment. Cite any AI-generated text you include and add a brief statement explaining which tools you used and how.
Adapted from the published examples below and the AI Assessment Scale.
Published statements for Level 3
Short excerpts from university teaching centers. Follow the links for the full statements.
“It is important to critically assess and cite any AI-generated content used in your assignments and projects.”
“Your use of AI tools must be documented and cited to conform to this course’s expectations.”
“Place any AI-generated text in quotation marks with attribution.”
Mixing levels in one course
Many published templates combine a default with exceptions set per assignment. Carnegie Mellon’s example states that permitted AI use will be named in the assignment prompt, and otherwise the default is that generative AI is disallowed. Recent syllabus research finds instructors moving the same way, from course-wide bans to task-based rules (Chirikov, 2026).
What you get
Pick a level and answer a few questions. The builder writes a student-facing statement you can paste into an assignment, plus private notes for your planning.
Go to the builder ↓For this essay, you may use AI to brainstorm topics and find sources to check. You must write the argument and all drafts yourself. AI may not draft or rewrite your text. At the end, add two or three sentences saying which tool you used and how. Verify every source AI suggests against the original. Do not paste classmates’ work or personal information into AI tools. If you are unsure whether a use is allowed, ask me first.
About the levels. Level names follow the AI Assessment Scale (Perkins, Roe & Furze). The levels describe conditions of AI use, not rising educational quality: a higher level is not a better assignment. Level 4 allows AI throughout a task with a set goal. Level 5 is for tasks designed around exploring what AI can do.
Full AI
AI may be used throughout the task. Students direct the tools, verify results and remain accountable for the final work.
Works well for: professional practice tasks, advanced projects and assignments where using AI well is part of the skill.
Draft syllabus statement
You may use generative AI throughout this assignment. You remain fully responsible for the accuracy, originality and quality of what you submit. Disclose the tools you used and describe how they shaped your work. Do not enter personal, confidential or other students’ information into AI tools.
Adapted from the published examples below and the AI Assessment Scale.
Published statements for Level 4
Short excerpts from university teaching centers. Follow the links for the full statements.
“You are welcome to use generative AI programs (ChatGPT, DALL-E, etc.) in this course.”
“You may use AI tools… for any part of the assignment process, including brainstorming, research, writing, and editing.”
“You must inform the instructor when you have used an AI tool.”
Mixing levels in one course
Many published templates combine a default with exceptions set per assignment. Carnegie Mellon’s example states that permitted AI use will be named in the assignment prompt, and otherwise the default is that generative AI is disallowed. Recent syllabus research finds instructors moving the same way, from course-wide bans to task-based rules (Chirikov, 2026).
What you get
Pick a level and answer a few questions. The builder writes a student-facing statement you can paste into an assignment, plus private notes for your planning.
Go to the builder ↓For this essay, you may use AI to brainstorm topics and find sources to check. You must write the argument and all drafts yourself. AI may not draft or rewrite your text. At the end, add two or three sentences saying which tool you used and how. Verify every source AI suggests against the original. Do not paste classmates’ work or personal information into AI tools. If you are unsure whether a use is allowed, ask me first.
About the levels. Level names follow the AI Assessment Scale (Perkins, Roe & Furze). The levels describe conditions of AI use, not rising educational quality: a higher level is not a better assignment. Level 4 allows AI throughout a task with a set goal. Level 5 is for tasks designed around exploring what AI can do.
AI Exploration
The task is about AI itself. Students experiment with AI tools, test them or critique them.
Works well for: AI-literacy units, design and computing courses, and assignments that evaluate AI itself.
Draft syllabus statement
This assignment asks you to explore generative AI as part of the task itself. Experiment, compare outputs and critique what the tools do well and poorly. Document your process, including prompts and decisions, and reflect on what you learned about the tools and the subject.
Adapted from the published examples below and the AI Assessment Scale.
Published statements for Level 5
We found few fully exploratory statements. These excerpts come from courses that treat AI as a professional tool or ask students to question it.
“This course treats generative AI as a professional tool you’ll increasingly encounter in academic and workplace settings.”
“I must always apply critical thinking to anything ChatGPT (or other AI models) tell me.”
“Students are encouraged to explore and utilize generative AI tools… as part of their learning and coursework.”
Mixing levels in one course
Many published templates combine a default with exceptions set per assignment. Carnegie Mellon’s example states that permitted AI use will be named in the assignment prompt, and otherwise the default is that generative AI is disallowed. Recent syllabus research finds instructors moving the same way, from course-wide bans to task-based rules (Chirikov, 2026).
What each level allows
Based on the AI Assessment Scale’s level descriptions. Adjust the details for each assignment.
Brainstorm ideas
Find and summarize sources
Outline or plan
Draft text or code
Edit and proofread
Include AI output in final work
Experiment with or critique AI as the task
Disclose AI use
Worked examples by discipline
How the levels map to real assignments. These are illustrations to adapt, not findings from a study.
Medication dosage calculation quiz
Students must show they can calculate doses unaided. Patient safety depends on it.
Evidence of learning: Supervised, in-class quiz.
Proof-writing exam
The skill being assessed is constructing the argument yourself.
Evidence of learning: Handwritten work in class, with a short oral follow-up if needed.
Primary-source research paper
AI can suggest search terms and help plan, but the analysis of sources is the learning goal.
Evidence of learning: Annotated bibliography and a draft submitted before the final paper.
Lab report
AI can help find background literature. Data analysis and interpretation stay with the student.
Evidence of learning: Lab notebook and raw data submitted with the report.
Argumentative essay with revision memo
Students draft with AI feedback and decide which suggestions to keep.
Evidence of learning: A memo explaining what AI suggested, what they kept and why.
Market analysis report
Mirrors workplace practice, where analysts draft with AI and verify the numbers.
Evidence of learning: Source list with each figure checked against the original.
Design project report
AI can help draft sections and check routine calculations. Design choices and safety judgments stay with the student.
Evidence of learning: A design review where students defend their choices and show key calculations by hand.
Capstone software project
Professional developers use AI coding assistants. The design and testing decisions are what you grade.
Evidence of learning: Code walkthrough and a test plan the student explains live.
Critique of an AI-generated lesson plan
Evaluating AI output is the skill: spotting errors, bias and weak pedagogy.
Evidence of learning: Annotated critique and a revised plan.
Build a full course policy
A level alone does not settle whether AI use is appropriate. Start with the learning objective, name what AI may do, and say what students must still do and defend themselves. The builder writes a student-facing policy and a rationale for you.