Building an AI-Use Appendix for Assignments That Require Transparency

An AI-use appendix is a short, honest section, usually placed after your references, that names the tools you used, the specific tasks you used them for, and what stayed your own work. Here’s a ready-to-adapt AI use appendix assignment example, plus the four pieces every good one includes and the wording that keeps it factual instead of defensive. More courses now ask for exactly this, and even when yours doesn’t, a clean appendix is one of the cheapest forms of protection you can add to a paper.

The goal isn’t to write a confession or a legal disclaimer. It’s to make your process legible, so an examiner never has to guess how the work came together.

Key takeaways

• An AI-use appendix documents how you used AI: which tools, which tasks, what you did with the output, and what remained yours.
• Be concrete. “Used ChatGPT to suggest counterarguments in paragraph 4” beats “used AI for help.”
• Disclosing allowed use reads as diligence, not guilt; undisclosed use discovered later is the real risk.
• Follow your course’s required format when there is one; otherwise a short paragraph or bulleted list works.
• An appendix doesn’t stop a detector from flagging you, so pair it with version history and notes.

Why an appendix, and why now

Instructors are caught in the same bind as students. They can’t reliably tell how much AI went into a submission, detectors are unreliable, and banning tools outright hasn’t worked. So a growing number have landed on transparency as the compromise: use AI within the rules, but *tell me how*. The AI-use appendix is the artifact that satisfies that ask.

For you, it does something valuable even when it isn’t required. It moves you from a defensive posture (“prove I didn’t cheat”) to a documented one (“here’s exactly what I did”). Those are very different places to stand if a question ever comes up. An appendix is a close cousin of a full AI disclosure statement; the appendix just lives on the assignment itself.

The four things every appendix should cover

You don’t need a long document. You need four elements, stated plainly.

1. The tools. Name each one specifically: ChatGPT, Claude, Grammarly, Copilot, whatever you actually used. “AI” alone is too vague.
2. The tasks. What did you use each tool *for*? Brainstorming, outlining, grammar checking, explaining a concept, suggesting counterarguments. Tie it to specific parts of the work when you can.
3. What you did with the output. Did you rewrite suggestions in your own words? Discard most of them? Verify a claim independently? This is where you show the tool informed your thinking without authoring your prose.
4. What remained yours. A clear statement of the parts that are entirely your own: the analysis, the argument, the writing.

Cover those four and an examiner has everything they need.

An AI use appendix assignment example to adapt

Here’s a complete AI-use appendix for a typical humanities essay where AI was allowed for limited support. Copy the structure and swap in your details.

> Appendix: AI Use Statement > > Tools used: ChatGPT (GPT-4), Grammarly. > > How I used them: > – *ChatGPT* — I asked it to explain the difference between two theoretical frameworks I was confused about, and later to list common objections to my thesis. It suggested three objections; I researched the strongest one further and addressed it in paragraph 5. I did not use any of its wording in the essay. > – *Grammarly* — Used for spelling and grammar checks on my finished draft. > > What is my own work: The argument, structure, analysis, and all prose in this essay are mine. I wrote a full rough draft before using either tool. Draft history is available on request.

Notice the tone: matter-of-fact, specific, and complete. It doesn’t apologize, and it doesn’t oversell. That’s the register to aim for.

A shorter version for light use

If you barely used AI, you don’t need the full treatment. A single sentence does it:

> AI use: I used Grammarly for spelling and grammar only. No generative AI was used to produce any content in this assignment.

Or, for a small allowed use:

> AI use: I used ChatGPT once to clarify a concept from the reading; all analysis and writing are my own.

Short, honest, done.

Wording that keeps it factual, not defensive

The difference between an appendix that helps you and one that makes you look nervous is almost entirely tone. A few principles:

• State, don’t justify. “I used ChatGPT to suggest counterarguments” is a fact. “I only used ChatGPT a tiny bit and I promise all the ideas are really mine” sounds anxious. Facts read as confidence.
• Be specific. Vagueness invites suspicion; precision closes it. Name the task and, where possible, the paragraph.
• Don’t over-disclose the trivial as if it were significant. Spell-check is spell-check; you can note it in a line without ceremony.
• Match the language to your policy. If your course uses a “AI-assisted vs. AI-generated” framework, use their terms. If they ask you to declare a percentage or a level, follow that format exactly. This guide on when and how to tell your professor you used AI covers matching your disclosure to different policy styles.

The thing an appendix can’t do

Be clear-eyed about one limit: an AI-use appendix does not stop a detector from flagging your work. The two are independent. A detector scans your prose for machine-like statistical patterns and returns a score regardless of what your appendix says. It doesn’t read your disclosure and adjust.

And detectors flag honest writing more than you’d hope. OpenAI shut down its own AI Text Classifier in July 2023, citing low accuracy. A 2023 study in *Patterns* by Stanford’s Weixin Liang and James Zou found seven detectors flagged more than half of 91 human-written TOEFL essays as AI, with one flagging nearly all of them. So even a fully self-written essay with a clean appendix can draw a suspicious score.

What the appendix *does* provide is context that helps a human interpret that flag correctly. Paired with version history and notes, it lets you say: here’s my honest account of my process, here’s the draft evolving over time, and here’s why a probability score isn’t the whole story. That combination is far stronger than any one piece. It’s also worth running your finished work through a detector yourself to see how your writing reads to a detector before you submit, so nothing surprises you.

Frequently asked questions

What is an AI-use appendix on an assignment? It’s a short, honest section, usually after your references, that documents how you used AI: which tools, which tasks, and what you did with the output. Instead of an examiner wondering whether AI was involved, they can read exactly what happened. Some instructors require it; others value it as good faith even when it isn’t mandatory.

What should I include in an AI-use appendix? Four things: which tools you used, the specific tasks each was used for, how you handled the output, and what remained your own work. Be concrete, “used ChatGPT to suggest counterarguments in paragraph 4” beats “used AI for help.” Follow your course’s format if required; otherwise a short paragraph or bulleted list is plenty.

Does writing an AI-use appendix make me look guilty of cheating? No, when the use was allowed. A factual disclosure of legitimate assistance reads as diligence, not a confession. What looks bad is undisclosed AI use that surfaces later. An appendix puts your process on the record on your terms, which is the strongest position if anyone asks.

Do I still need an appendix if I only used AI a little? If your course requires disclosure, yes, because the requirement is about transparency, not amount. If it isn’t required, a short appendix for anything beyond basic spell-check is cheap insurance. When in doubt, disclose; a one-line note about small, allowed use costs nothing.

Will an AI-use appendix stop my work from being flagged by a detector? No, they operate independently. A detector scans your prose for machine-like patterns regardless of your appendix and can still return a score. The appendix provides context that helps a human interpret a flag correctly. Pair it with version history and notes for the strongest position.

The short version

An AI-use appendix is just an honest, specific account of how AI touched your work: the tools, the tasks, what you did with the output, and what stayed yours. Keep it factual and precise, follow your course’s format when there is one, and remember it documents rather than defends. It won’t stop a detector from misfiring, but combined with a process record, it puts you in the best possible spot: transparent, on the record, and easy to believe.