There is no single AI policy at Jefferson Academy Secondary. What counts as appropriate AI use depends on your course, your assignments, and the skills you are building, so the decision is yours. The AI committee's recommendation is that every syllabus state that decision in plain language before the first assignment, so students are not left guessing. Refer to the Student/Parent Handbook language Page 10
The options below are starting points, not a fixed menu. Mix and match language across sections, adjust wording to fit your course, and pair whichever approach you choose with the Generative AI Usage Menu if you want to spell out exactly which uses count as "supporting" or "replacing" for a given assignment. Teachers found last year's four options too coarse for how they actually wanted to run their classrooms, so this version adds tiers in between full restriction and open use, a disclosure clause, and a much wider set of consequence examples.
Whatever language you choose, the AI committee recommends covering four things:
The strictest tier: no AI at any stage, for any reason, on any course activity. Use this when the entire point of the class or the assignment is demonstrating what a student can do unaided — there is no partial-credit version of this rule.
Sample syllabus language
"This is a brain-only classroom. You are expected to complete all assignments, assessments, and in-class work using your own thinking, knowledge, and skills. The use of AI tools—including but not limited to ChatGPT, other text generators, or AI-assisted problem solvers—is not permitted for any course activity. All work submitted should reflect the student's own knowledge, expression, and skills."
Supporting tier — student does the thinking
AI tools may be used only as a study partner or learning support.
Sample syllabus language
"AI tools may be used only as a study partner in this course, not as a stand-in for your own thinking. You can ask AI to explain a concept you're stuck on, walk through a similar example, or generate extra practice problems to test yourself before a quiz. What you can't do is ask it to produce answers, draft any part of an assignment, or work through a problem that you then copy over as your own. Every assignment you submit needs to come from your own reasoning and the skills you're building, not a shortcut around either."
Supporting tier — student creates, AI critiques
This tier sits between Study Partner and full editing help. AI may respond to work you have already produced in your own words, but it may not generate replacement text.
This was one of the gaps teachers ran into last year: wanting to let AI critique a draft without also opening the door to AI writing it.
Sample syllabus language
"AI tools may be used only to give you feedback on work you have already written yourself — pointing out where an argument is weak, where a step likely has an error, or where a transition doesn't hold together. AI may not rewrite your sentences, fix the error itself, or hand you replacement text to paste in. The revisions in your final submission need to be yours, made because of what the feedback pointed out, not written by the tool that pointed it out."
For teachers whose expectations shift constantly. Rather than restating AI rules in every assignment, the syllabus establishes a standing three-color system once, and each assignment simply carries a label — in the title or stamped on the handout — so students never have to guess. The colors themselves are defined here, in the syllabus, not reinvented per assignment:
Green — AI Open
Use AI however it helps. Disclosure encouraged, not required.
Yellow — AI With Limits
AI allowed for the steps named on the assignment, and only those. Disclosure required.
Red — Brain Only
No AI at any stage, including planning and editing.
Sample syllabus language
"Each assignment in this course is labeled Green, Yellow, or Red for AI use. If an assignment has no label, treat it as Red."
This tier splits the rule by location rather than by task: what happens in the room stays unaided, while work done outside class has more room for AI support. It works well for courses where in-class time is already built around live, observed performance.
Sample syllabus language
"Anything produced during class time is brain-only: no AI on in-class writing, problem sets, labs, or assessments. Outside of class you may use AI to study, to prepare, and to get feedback on your own work. Because your in-class work is where you show what you can do unaided, using AI outside class in a way that leaves you unable to do the same work in class will show up quickly."
The most permissive tier: no restriction on how AI is used, paired with full ownership of the result. This works for courses where the skill being assessed is judgment and verification rather than the underlying production task itself.
Sample syllabus language
"AI tools are permitted in this course without restriction. That permission comes with a condition: you are accountable for every claim, citation, number, and line of code you submit. AI output that is wrong is your error, not the tool's. Expect to be asked to explain any part of your work in conversation."
Teachers who otherwise restrict AI often add a carve-out so students using AI as an access tool are not caught by the rule.
Sample syllabus language
"Regardless of the AI expectations for a given assignment, you may always use assistive tools — including AI-based text-to-speech, speech-to-text, translation, and reading support — to access the material and to communicate your own thinking. Tell me what you are using so I can plan around it. This does not extend to generating the content of your work."
Whenever a tier above Brain-Only applies, the AI committee recommends that students include a short process note with their submission: which tool was used, what it was used for, and how the output was incorporated into the final work. This isn't a schoolwide mandate — it's a practice you can adopt to whatever degree fits your course. It mirrors the AI committee's AI Documentation guidance — teachers can recommend the short version (tool, purpose, integration) or the fuller reflection statement, depending on the assignment. Add one line to your syllabus naming which version you'd like to see, so students aren't left guessing between a citation, an appendix, and a reflection paragraph.
Each version below asks for a different amount of work from the student. Copy the one that matches your assignment into your syllabus or onto the assignment sheet.
1. Citation only
If you used an AI tool on this assignment, please add a citation line at the end naming the tool, the version if you know it, a link to the chat if you have one, and the date: "Google. Gemini. [link to chat]. Accessed 12 Sept 2026." Nothing further is needed.
2. Citation with purpose
Add one line at the end of your work naming the tool, what you used it for, and a link to the chat if you have one: "I used Gemini to generate practice questions while I studied for this unit. [link to chat]" One sentence is enough. If you used more than one tool, name each.
3. Appendix
If you'd like, attach your AI work as an appendix after the final page: a link to the chat, or a copy of the prompts you wrote and the responses you received. The appendix isn't graded for quality, but a complete one gives me something to reference if a question comes up later.
4. Reflection paragraph
Consider ending your submission with a short paragraph (3–5 sentences) answering three questions: What did you ask AI to do? What did it give you that you kept, and what did you throw out? What would your work look like if you had not used it? Whether this counts toward your grade is noted on the assignment.
5. Citation and reflection together
Include both an AI citation line (tool, version, date, and a link to the chat if you have one) and a short reflection paragraph describing what you asked for, what you used, and what you changed. The citation gives me a record — and a way to check it — that AI was involved; the reflection shows me that the thinking was still yours.
6. Process log across a project
For multi-day work, consider keeping a running log with your project: each time you use AI, add a dated line with the tool, a link to the chat if you have one, what you asked, and what you did with the answer. Share the log with the final product. This can stand in for a single disclosure statement at the end.
7. In-text marking
Where AI contributed directly to something you kept — a phrase, a line of code, a structural choice — consider marking it in the draft you submit (highlight, comment, or footnote). Your final clean copy stays unmarked; the marked draft comes with it.
8. Verbal disclosure
No written statement is needed. Be ready to tell me, if I ask, what role AI played in your work and to walk me through any part of it. Not being able to explain your own submission is the concern, not the AI use itself.
9. No disclosure required
AI use on this assignment does not need to be reported. You are still responsible for everything you submit being accurate and your own to defend. Worth stating explicitly if your other assignments do require disclosure — otherwise students assume the stricter rule.
Before any grade consequence, follow the AI committee's Suspected Inappropriate AI Use protocol: a conversation with the student, framed around understanding their process rather than an accusation, comes first. Grade consequences apply once that conversation and any documented evidence confirm misuse.
Teachers here set consequences at very different levels, from a no-penalty redo to an automatic zero. The AI committee recommends stating whichever approach you choose in the syllabus in advance rather than deciding after the fact — either end of the spectrum is defensible as long as students know before it happens. Pick the language below that matches how you'd actually respond, or use it as a starting point to write your own.
Rewards catching yourself before you're caught
Consider setting your own figure to match your department's academic integrity practice — the 10% reduction below is a starting point, not a fixed number. What matters more is keeping the self-report incentive, since it's the part that surfaces problems you'd otherwise never hear about.
Sample syllabus language
"If you self-report inappropriate use of AI on an assignment—either before or after being questioned—you will have the opportunity to redo the assignment with a 10% grade reduction. If you do not self-report and inappropriate AI use is identified, the assignment will receive a grade of 0 and cannot be redone."
Strictest of the common options
Several teachers choose this deliberately. It holds up well as long as the syllabus states it plainly on day one and the documentation steps were followed first.
Sample syllabus language
"Inappropriate use of AI on an assignment results in a grade of 0 for that assignment. The assignment may not be redone or replaced, and the zero is included in the term grade. A guardian contact and an academic integrity referral accompany the zero."
Automatic zero, one recovery
This tier keeps the automatic zero but gives students one way to earn back some credit, capped below full value so the redo doesn't erase the consequence entirely.
Sample syllabus language
"Inappropriate AI use results in a 0 on the assignment. You may replace that 0 by completing a written reflection on what you were supposed to be practicing and then redoing the assignment under supervision, for a maximum of 80%. This option is available once per semester."
Work redone, first score doesn't count
This tier treats the original submission as void rather than zeroed — the redo becomes the grade of record on the first offense, with a stricter consequence held in reserve for repeat instances.
Sample syllabus language
"If inappropriate AI use is identified, the submitted work does not count. You will redo the assignment under supervised conditions, and the redo is the grade of record. Repeat instances move to a zero with no redo."
First instance instructional
This ladder treats a first AI-misuse incident as a teaching moment and reserves the harshest consequence for repeat offenses, so students have a clear sense of how the response escalates.
Sample syllabus language
"First instance: a conversation with me, and you redo the assignment for full credit.
Second instance: you redo the assignment for a maximum of 70%, and I contact your guardian.
Third instance: the assignment receives a 0 with no redo, and I submit an academic integrity referral."
No grade penalty by itself
This tier treats the incident as a gap in evidence rather than a punishment opportunity — the goal is confirming what the student can actually do, not penalizing the attempt.
Sample syllabus language
"Inappropriate AI use means I do not yet have evidence of what you can do. The submitted work is set aside and you will demonstrate the same skill in a different format—an oral explanation, a timed in-class task, or a conference with me. That demonstration becomes the grade."
These examples are built for a single assignment sheet, rubric, or project brief rather than the syllabus itself. Use them when a task needs something more specific than your course-wide policy, or when a particular presentation, paper, or project calls for its own AI use statement.
A one-line tag at the top of the sheet
Spells out exactly what's allowed rather than naming a category — students won't have the class AI use guide open, so the stamp needs to stand on its own. Pick the version that matches this assignment.
Sample assignment language — Brain Only
"AI Use for This Assignment: Not permitted. Do not use AI at any point while completing this assignment, including for brainstorming, outlining, or checking your work."
Sample assignment language — Supporting
"AI Use for This Assignment: Gemini may be used to help you get started (for example, generating an outline or a list of ideas) or to give feedback on a draft you already wrote. Gemini or other AI may not write your first draft or hand you text to copy into your final work."
Sample assignment language — Replacing
"AI Use for This Assignment: Gemini may be used to produce parts of your submission directly (for example, drafting a section or generating an image), as long as you review, edit, and take responsibility for anything Gemini produces before you submit it."
Permission changes by stage of work
For assignments with distinct stages — research, drafting, revision — where Gemini use should change depending on where a student is in the process. Fill in each stage before handing this out; the stage names can be renamed to match a presentation's build process, a paper's drafting process, or a project's production timeline.
Sample assignment language
"Use of Gemini on this assignment depends on the stage of your work. Research/Planning: [open use / permitted / not permitted]. Drafting/Building: [open use / permitted / not permitted]. Revision/Polishing: [open use / permitted / not permitted]."
Concrete uses named for this task
A specific list rather than a category label — spells out exactly what's allowed and what isn't for this particular deliverable, whether it's a paper, a presentation, or a project.
Sample assignment language
"Gemini Can: [specific allowed use], [specific allowed use]. Gemini or other AI tool Cannot: [specific disallowed use], [specific disallowed use]. Anything not listed above defaults to [allowed / not allowed] for this assignment."
Assessed alongside everything else
Puts the AI expectation directly into the grading rubric instead of as a separate policy statement, so it's part of the score rather than a rule enforced on the side.
Sample assignment language
"Original Thinking & Authentic Voice — 4: Work reflects independent reasoning; any Gemini use stayed within the uses named for this assignment. 3: Mostly independent, with minor reliance on Gemini beyond what was named. 2: Significant sections reflect Gemini output rather than the student's own reasoning or voice. 1: Submission is substantially Gemini-generated with little evidence of the student's own work."
Tied to what the task is measuring
Frames the rule around what the assignment is actually assessing, keeping "who's doing the thinking?" tied directly to this specific task rather than a general policy.
Sample assignment language
"This assignment measures your ability to [skill being assessed — e.g., organize an argument, analyze data, design a solution]. Gemini may help you with [permitted supporting task], but the skill above must be done brain-only. If Gemini touches the part being measured, it isn't evidence of what you can do."
One frame, any subject
A fill-in-the-blank frame any department can adapt to its own task type — an essay, a lab report, a problem set, a presentation, a project — rather than a fixed example tied to one subject.
Sample assignment language
"For this [essay / lab report / problem set / presentation / project], Gemini may be used to [permitted use], but may not be used to [prohibited use]. No other AI Tools may be used. This follows the course's general AI use expectations, adjusted for what this task is asking you to show."
Says exactly where the evidence goes
Distinct from the course-wide Disclosure Recommendation above: a short, task-specific instruction pointing at exactly where the evidence belongs on this particular deliverable.
Sample assignment language
"If Gemini was used anywhere in producing this [assignment / presentation / paper / project], note where and how on [the title slide / the first page / an appendix]: what you asked Gemini to do, what you kept, and a link to the chat if you have one. Gemini is the only permitted AI Tool. No otehr AI tool may be used."
Disclosure stays individual
For collaborative work, where the deliverable is shared but disclosure needs to stay individual so one member's use doesn't get attributed to the whole team — or hide behind it.
Sample assignment language
"Because this is a group [project / presentation / paper], each team member discloses their own Gemini use individually, even though the final product is submitted as one file. Include a line per person: name, whether Gemini was used, and what it was used for. One member's disclosure doesn't cover the team."
Permission set per assignment
Permissible use of AI will vary by assignment. Students must follow the teacher's guidance on whether AI is allowed and how it may be used. Unless an assignment explicitly states that AI use is allowed, assume it is a "brain-only" assignment—completed entirely with your own knowledge, expression, and skills.
Last year this tier only told students to "check the assignment," without giving teachers a quick way to say what's actually allowed. Attach a short checklist to the assignment sheet instead of writing new policy language each time:
For this assignment, AI may be used for:
Anything left unchecked is off-limits for this assignment. See the Generative AI Usage Menu for what each of these looks like in practice.
Partial replacing, named explicitly
Some courses want AI available for the early, exploratory stages of a project (finding sources, generating options, mapping an approach) while keeping the actual writing, analysis, or problem-solving entirely the student's own. Name exactly which uses are in bounds so students aren't guessing where the line falls:
In this course, AI may be used to:
Checked boxes above are examples only — set your own for your class. Every source or idea an AI tool contributes must still be evaluated, verified, and cited by the student; the final product must show your own synthesis, not the AI's.
AI detection tools are not reliable enough to be the basis for a consequence on their own. The AI committee's recommendation is to start from the work itself: unexplained shifts in voice, claims the student cannot talk through, sources that do not exist, or a gap between this submission and everything else the student has produced.
See Suspected Inappropriate Use for the conversation and documentation steps, and Detecting AI Writing for what the signals can and cannot tell you.