How to Disclose AI Use in Academic Work
An AI disclosure statement tells a reader that you used an AI tool, explains what the tool did, and makes the boundaries of your own work visible. It can be as short as one or two sentences for a routine assignment, or more detailed for a thesis, dissertation, research project, code submission, or assessment that requires a prompt record.
There is no single disclosure rule that applies to every university or assignment. University of Washington guidance on assignment-level AI policies shows why course instructions matter: instructors may permit specific uses, prohibit others, and require students to identify the tool and explain how AI output was evaluated, modified, and verified. Tufts University’s 2026 AI guidelines similarly emphasize transparency while making disclosure expectations proportional to context.
For the broader framework behind permission, authorship, attribution, verification, and responsible assistance, use EssayEco’s academic integrity and responsible AI use guide.
| Core rule: Disclosure does not turn prohibited AI use into acceptable AI use. Check permission first, then disclose accurately, and cite AI-generated material separately when your style guide or instructor requires it. |
What Is an AI Disclosure Statement?
An AI disclosure statement is a short account of how artificial intelligence contributed to your academic work. Its purpose is transparency. It helps an instructor, supervisor, collaborator, or reader distinguish your own intellectual work from assistance provided by a generative AI system.
A strong statement usually answers four questions: Which tool did you use? What did you use it for? How much did it influence the submitted work? and What did you personally check, change, or decide? Some institutions also require the model or version, date of use, prompts, outputs, or an appendix.
Disclosure, permission, citation, and documentation are different
| Term | Main Question | What It Does |
| Permission | Was this use allowed? | Comes from the assignment brief, syllabus, instructor, department, or institutional policy. |
| Disclosure | How did AI contribute? | Explains the tool, purpose, scope, and human oversight behind the submitted work. |
| Citation | Am I using AI-generated words, ideas, images, or output as a source? | Credits material according to the required citation style or course rules. |
| Documentation | Can I show what I actually did? | Preserves prompts, outputs, dates, revisions, verification notes, or other records when required. |
If you need citation formats, use EssayEco’s guide on how to cite ChatGPT and other AI tools in academic work. Citation and disclosure can overlap, but one does not automatically replace the other.
When Should You Disclose AI Use?
Start with the most specific rule available. Assignment instructions generally matter more than broad assumptions about what universities “usually” allow. Emory’s responsible AI guidance recommends context-specific, proportionate disclosure and treats substantive AI contribution differently from minor stylistic assistance.
| AI Use | Disclosure Usually Needed? | Why / What to Check |
| Generating or reorganizing core ideas | Often yes | The tool influenced the intellectual structure or direction of the work. |
| Drafting paragraphs, summaries, explanations, or conclusions | Often yes | AI contributed wording or substantive content that may need both disclosure and citation. |
| Brainstorming questions or possible approaches | Depends on course rules | Even when no AI text is submitted, the tool may have shaped the thinking process. |
| Feedback on a draft or argument | Depends on course rules | Disclose if the assignment requires AI-use statements or if feedback substantially shaped revisions. |
| Coding, data analysis, translation, or image generation | Often yes | AI may have affected methods, outputs, interpretation, or creative material. |
| Basic spelling, grammar, or formatting assistance | Policy dependent | Some policies exempt minor embedded tools; others require any generative AI use to be acknowledged. |
| No AI use | Only if required | Some assignments ask every student to submit either a use statement or a truthful no-use declaration. |
When the instructions are silent and the use is more than trivial, the safest academic practice is to ask the instructor before submission rather than invent your own rule. The important distinction is between transparent documentation and retroactive justification.
What Should an AI Disclosure Statement Include?
1. Name the tool clearly
Identify the product you actually used, such as ChatGPT, Microsoft Copilot, Gemini, Claude, Grammarly with generative features, GitHub Copilot, or an institution-specific AI system. If the course requires the model, version, URL, or date, include those details too.
2. State the specific purpose
Avoid vague wording such as “AI helped with the assignment.” Say whether the tool was used to brainstorm questions, test an outline, explain a concept, generate code alternatives, improve sentence clarity, translate text, create an image, or critique a draft.
3. Describe the scope of AI involvement
Make the boundary visible. Did AI influence only the planning stage? Were several sentences revised after AI feedback? Did you generate code and then test it independently? Did AI suggest sources that you later located and evaluated yourself? Scope matters because “used AI” can describe very different levels of involvement.
4. Explain your human review and verification
State what you personally checked. For factual work, that may include confirming claims against credible sources, testing calculations, reading the original literature, checking code, or correcting fabricated citations. EssayEco’s guides on verifying AI-generated references and AI hallucinations in academic research cover those verification steps in depth.
5. Add any documentation your course requires
Some instructors ask for prompts, outputs, screenshots, an appendix, a process log, or a share link. Others want only a brief note. Do not attach a large prompt transcript unless the assignment requests it, and do not include confidential, personal, proprietary, copyrighted, or research-sensitive information simply to prove that AI was used.
How to Write an AI Disclosure Statement in 7 Steps
Step 1: Check the rules before using the tool
Read the assignment brief, syllabus, academic-integrity policy, departmental guidance, and any instructor announcements. Note which AI uses are allowed, restricted, or prohibited, and whether a disclosure format is specified.
Step 2: Keep a simple record while you work
Write down the tool, approximate date, task, and purpose when you use AI. This takes seconds and is much easier than reconstructing your process after the paper is finished.
Step 3: Separate assistance from authorship
Identify what the tool actually contributed and what remained yours. Planning assistance, wording suggestions, source discovery, coding support, and substantive drafting should not be collapsed into the same description.
Step 4: State the extent of the contribution
Use concrete language: “used to generate five possible research questions,” “used to identify weaknesses in my outline,” or “used to suggest alternative Python syntax that I tested and revised.” Avoid percentages unless you have a meaningful and verifiable way to calculate them.
Step 5: Describe verification and revision
Explain how you checked the output rather than implying that AI was automatically reliable. Mention source verification, comparison with course materials, manual calculations, code testing, fact checking, or your own substantive revisions when these actions occurred.
Step 6: Add citation or supporting records if required
If you quote, paraphrase, reproduce, or closely rely on AI output, follow the required citation guidance. If the instructor asks for prompts or transcripts, provide them in the specified format. Disclosure alone is not a substitute for required attribution.
Step 7: Place the statement where the instructor expects it
Use the designated cover-sheet field, author note, acknowledgement, appendix, methodology section, footnote, end note, or submission form. If no location is specified, a concise note at the end of the assignment is often easy for an instructor to find, but local instructions always take precedence.
| A useful formula: Tool + purpose + scope + human review. Add model/version, date, prompts, output records, citations, or placement details only when the course or context requires them. |
AI Disclosure Statement Examples for University Assignments
The examples below show how the same structure can be adapted to different uses. They assume the described AI use was permitted. Do not copy a sample that understates what you actually did.
| Use Case | Example Disclosure |
| Brainstorming | “I used ChatGPT to generate possible angles for the topic and compare several research-question options. I selected and revised the final question independently and did not use AI-generated prose in the submitted paper.” |
| Outline feedback | “I used Microsoft Copilot to review an early outline for gaps and possible counterarguments. I evaluated the suggestions against the assignment criteria and created the final structure and argument myself.” |
| Language revision | “I used an AI-enabled writing tool to suggest sentence-level clarity and grammar revisions after drafting. I reviewed each change, retained only appropriate edits, and remained responsible for the final wording and meaning.” |
| Source discovery | “I used ChatGPT to suggest search terms and possible sources. I independently located the original publications, verified their bibliographic details, evaluated their credibility, and cited only sources I personally reviewed.” |
| Coding support | “I used GitHub Copilot to suggest alternative code syntax for data cleaning. I tested, debugged, and modified the code, checked the outputs against the raw data, and take responsibility for the final analysis.” |
| AI-generated image | “I used [tool] to generate the illustrative image shown in Figure 2 from a prompt I created. I revised the prompt and selected the final image; the figure is labeled as AI-generated in accordance with the assignment requirements.” |
Example for a thesis, dissertation, or research project
UMass Amherst’s graduate-student AI policy is one example of a stricter research context: it requires an AI disclosure statement for degree-related research and asks students to state the tool names, specific purposes, and extent of human oversight. A thesis or dissertation statement may therefore need more detail than a routine course assignment.
Example: “During the literature-mapping stage, I used [tool/model] to generate search-term combinations and to identify possible themes. I did not rely on AI summaries as evidence. All cited sources were located in the original databases, read, evaluated, and cited by the author. AI was not used to generate the findings or final analysis.”
Where Should You Put an AI Disclosure Statement?
| Location | Best Fit | Notes |
| Assignment cover sheet / submission form | When the course provides a dedicated field | Use the exact format and character limit provided. |
| End note or acknowledgement | Short course assignments | Works well for a brief process statement that is separate from the main argument. |
| Author note | Formal academic papers when requested | Useful when the instructor or style guide specifies an author-note disclosure. |
| Methods / methodology section | Research projects where AI affected a method | Describe AI involvement where readers need it to understand or reproduce the process. |
| Appendix | Prompt logs, outputs, or detailed records | Use when documentation is required but would interrupt the main text. |
| Figure / table note | AI-generated or materially AI-modified visuals | Place disclosure close to the affected content when required. |
Do not hide a required disclosure in a place the instructor is unlikely to see. The location is part of compliance, not just formatting.
AI Disclosure vs. AI Citation: Do You Need Both?
Sometimes. Disclosure describes your process; citation attributes material. If AI generated text, an image, code, or other output that you quote, reproduce, paraphrase, or treat as a source, your instructor or style guide may require a citation in addition to a disclosure statement. If AI was used only for brainstorming or feedback, there may be nothing to cite even though disclosure is still required.
This is why the sequence matters: first determine whether the use was permitted when using AI for university assignments responsibly; then disclose the process; then apply the appropriate citation rule to any AI-derived content.
Does Disclosing AI Use Make It Acceptable?
No. Transparency is not permission. If an assignment prohibits generative AI, writing “I used ChatGPT to draft the conclusion” does not cure the violation. Likewise, a disclosure cannot make fabricated evidence, invented references, confidential-data exposure, unauthorized collaboration, or misrepresentation acceptable.
A disclosure also should not function as a disclaimer that transfers responsibility to the tool. You remain responsible for the accuracy, relevance, legality, ethics, and academic integrity of what you submit. Tufts explicitly frames accountability and verification as human responsibilities even when AI is involved.
Keep a Private AI Use Log
A lightweight process log helps you write an accurate disclosure and explain your workflow if questions arise. The log is for your records unless the assignment requires submission.
| Record | What to Note |
| Date / stage | When the AI interaction occurred and whether it was planning, drafting, revising, coding, analysis, or presentation. |
| Tool / model | Product name and model or version if visible and relevant. |
| Purpose | What you asked the tool to help with. |
| Prompt / input | A short description or the exact prompt if the course requires prompt-level documentation. |
| Output used | What, if anything, influenced the submitted work. |
| Your action | What you rejected, rewrote, tested, verified, or decided independently. |
| Evidence checked | Sources, calculations, data, code tests, or course materials used to verify the result. |
Do not preserve or submit sensitive information in a prompt log if doing so would violate privacy, confidentiality, research ethics, copyright, or data-handling rules. Documentation should support integrity without creating a second compliance problem.
Common AI Disclosure Mistakes
| Mistake | Why It Is Weak | Better Approach |
| “I used AI.” | Too vague to show what the tool contributed. | Name the tool and describe the specific task. |
| Using a disclosure to excuse prohibited use | Disclosure cannot override the assignment rules. | Check permission before using the tool. |
| Claiming “all sources were verified” without actually checking | Creates a false record of diligence. | Describe only verification steps you genuinely performed. |
| Listing the tool but not its role | The reader still cannot distinguish AI assistance from your work. | State purpose and scope. |
| Copying a generic disclosure template | May misrepresent your real workflow. | Adapt the statement to what actually happened. |
| Including sensitive prompt content unnecessarily | Can create privacy or confidentiality risks. | Provide only the documentation required and redact/omit restricted information appropriately. |
| Assuming citation replaces disclosure | Citation attributes material but may not explain the process. | Follow both requirements when both apply. |
A Quick Checklist Before You Submit
- I checked the assignment, syllabus, instructor guidance, and relevant academic-integrity policy.
- The AI use I am disclosing was permitted for this task.
- I named the tool accurately and included model/version/date details if required.
- I stated exactly what the AI tool was used to do.
- I described the scope of its contribution without minimizing or exaggerating it.
- I explained my own review, verification, testing, and revisions where relevant.
- I cited AI-generated material separately if the assignment or style guide requires it.
- I included prompts, outputs, screenshots, or an appendix only if required.
- I placed the disclosure where the instructor expects it.
- I did not expose confidential, personal, copyrighted, or research-sensitive information in the disclosure record.
Final Takeaway
A useful AI disclosure statement is specific enough to make your process understandable without turning into a transcript of every click. Name the tool, state the purpose, describe the scope, and explain the human review that actually occurred. Add citations, prompts, model details, dates, or appendices when the assignment requires them.
Most importantly, treat disclosure as the final transparency step in a larger academic-integrity workflow: verify that AI use is allowed, protect sensitive information, check AI outputs against credible academic sources, document meaningful assistance, and remain responsible for the work you submit.
Frequently Asked Questions
Do I need an AI disclosure statement if I only used ChatGPT for brainstorming?
Possibly. Some courses require disclosure of any generative AI use, while others focus on substantive contribution to the submitted work. If brainstorming was allowed but disclosure rules are unclear, check the assignment or ask the instructor. Do not assume that “no AI text was copied” automatically means “no disclosure is needed.” For a student-controlled ideation workflow, see how to use AI for brainstorming without replacing your own thinking.
Do I have to disclose Grammarly, spellcheck, or other writing tools?
It depends on the tool and the policy. Basic spelling and formatting tools may be treated differently from generative features that rewrite, expand, translate, or create content. Follow the rule for the feature you actually used, not just the product name.
Is an AI citation the same as an AI disclosure statement?
No. A citation attributes a source or output. A disclosure explains how AI contributed to your process. You may need one, both, or neither depending on the assignment, the type of AI use, and the required citation style.
Should I include every prompt I used?
Only if required. Many assignments need a concise statement rather than a full transcript. If prompt records are requested, follow the specified format and protect confidential or restricted information.
What if my assignment says nothing about AI?
Check the course syllabus and institutional or departmental rules. If those are also silent and your planned AI use is substantive, ask the instructor before relying on the tool. A short clarification can prevent a serious misunderstanding about unauthorized assistance.
Can I write a “no AI used” statement?
Yes, if the assignment requires a declaration and the statement is true. Do not make a no-use declaration if you used generative AI in a way covered by the course policy.
Will disclosing AI use protect me from an AI detector?
Disclosure is not a detector workaround. It is a record of your actual process. Recent higher-education research on student AI disclosure shows that disclosure decisions can be shaped by worries about judgment and stigma, which is another reason clear course policies and accurate process documentation matter.
