Is Using AI Plagiarism? How to Disclose It
Using AI is not automatically plagiarism, but passing off AI-generated text or ideas as your own work usually is, and so is using AI where your course or employer bans it. The rule that decides it is your institution's policy for that assignment. When in doubt, ask first, keep your drafts, and disclose what the tool did.
The short answer
Plagiarism means presenting work that is not yours as yours. Text a tool wrote for you, submitted as your own writing, fits that definition in most academic integrity codes, even though no human author is being copied. Using AI for brainstorming, explanations or grammar fixes is a different matter, and many courses allow it. There is no single global rule: your school, your instructor and, for publications, your journal set the line. This guide gives you a way to place any use on a spectrum, templates for disclosing it and the basics of citing AI tools. For the underlying definition, see what is plagiarism.
Why "is AI plagiarism?" has two answers
Traditional plagiarism is about taking from a source: another person's words, ideas or data. AI text raises two separate questions.
- Authorship. If a tool wrote the sentences, who is the author? Most policies say the student must be the author of submitted work, so unacknowledged generated text counts as misrepresenting authorship.
- Attribution. Language models produce text from patterns, and they can echo existing material or invent references that do not exist. Either way, you are responsible for what you submit.
Because of this, integrity offices often file undisclosed AI use under "unauthorised assistance" or "contract cheating" rather than classic plagiarism. The label matters less than the consequence: treat any undisclosed substantial AI use as a breach unless your policy clearly says otherwise.
A spectrum from safe to serious
Think of AI use as a ladder. The higher the rung, the more of the thinking and wording is no longer yours.
| Level | What you do | Usual status |
|---|---|---|
| 1. Spell and grammar fixes | Accept corrections to typos and punctuation | Almost always fine |
| 2. Explaining and brainstorming | Ask a tool to explain a concept or suggest angles, then close it and write | Often allowed, ask if unsure |
| 3. Outlining and feedback | Ask for a structure or for comments on your draft; you do the rewriting | Depends on the course; disclose |
| 4. Rewording your own draft | A paraphrasing or rewriting tool changes your sentences | Allowed in some courses; disclose and check meaning |
| 5. Drafting passages | A tool writes paragraphs that you edit | Usually restricted; needs explicit permission and attribution |
| 6. Ghost-writing | A tool writes the whole assignment | Treated as a serious integrity violation in most settings |
"Usually" and "often" are deliberate. Two instructors at the same university can set different rules for different assignments, and an exam may ban what a coursework project permits. Never assume that last term's permission carries over.
Where to find the actual rules
Do not rely on a blog post, a classmate or a tool's marketing for the rules. Look in these places, in this order:
- The assignment brief and syllabus. Many courses now state whether generative AI is allowed, restricted or banned, and what to disclose.
- Your institution's academic integrity policy. Search your university or school website for "academic integrity" and "generative AI".
- Your instructor or supervisor. A short email settles it and leaves a record (a script is below).
- For publications: the journal's author guidelines. Publishers commonly say that AI tools cannot be listed as authors and that their use should be disclosed. Check the specific journal rather than assuming.
- At work: your employer's AI and confidentiality policy. Pasting client or unreleased material into an online tool can breach confidentiality even when quality is not an issue.
AdeptWrite does not set rules for your institution, and none of this is legal advice. If the policy is silent, ask; "I did not know" carries little weight after the fact.
How to ask your instructor
A clear question gets a clear answer. Something like:
Hello Professor Adams, I am working on the case study due on 14 December. I would like to use an AI tool to check my grammar and to get feedback on the structure of my draft, but I will write all the content myself. Is that allowed, and would you like me to mention it in the submission? Thank you.
Keep the reply. If the answer is "no AI at all", that settles it for that assignment.
Disclosure statement templates
If your policy asks for disclosure, state what the tool did, what you did and how you checked the output. Keep it factual. Adapt these:
Light use (language support)
I used an AI grammar tool to correct spelling and punctuation in my final draft. All ideas, structure and wording are my own.
Brainstorming and feedback
I used ChatGPT to suggest possible angles for the topic and to comment on the clarity of my second draft. I did not copy any generated text into the submission. I wrote and revised the essay myself.
Rewording with a tool
I used a paraphrasing tool to reword three sentences in the discussion section. I compared each result with my original meaning and edited them. The sources were cited separately.
Generated material included (only where allowed)
The paragraph on page 4 was drafted with ChatGPT using the prompt "Summarise the main arguments for congestion pricing in two paragraphs." I edited the text and verified every claim against the sources cited in the reference list.
Put the statement where your course asks for it: a note after the conclusion, a cover sheet or a methods paragraph. Keep a copy of your prompts if you used generated text.
Citing AI tools in APA and MLA
When you quote or reproduce AI output, or when your style guide requires citing it, treat the tool as the source. The official guidance evolves, so confirm against the current APA and MLA pages or your department's handout. The formats below follow the structure those styles describe; dates are placeholders to replace with your own.
APA (7th edition)
- In-text: (OpenAI, 2025)
- Reference list: OpenAI. (2025). ChatGPT (Month Day version) [Large language model]. https://chatgpt.com
APA also advises describing the prompt in your text and noting that the output is not retrievable by other readers. Put long outputs in an appendix.
MLA (9th edition)
- In-text: ("Summarise the main arguments" )
- Works Cited: "Summarise the main arguments for congestion pricing in two paragraphs" prompt. ChatGPT, version or date, OpenAI, 5 Nov. 2026, chatgpt.com.
MLA suggests describing the prompt in the title position, naming the tool as the container, the company as publisher and the date you generated the text. For a fuller walkthrough, read how to cite ChatGPT, APA 7 citation guide and MLA format. The Citation Generator formats conventional sources such as books, articles and websites in APA 7, MLA 9, Chicago, Harvard and IEEE; for AI tools, always check the result against your style guide.
Citing the tool does not make the text yours. It shows where the words came from, and your policy still decides whether using them is allowed.
Verify everything the AI gives you
Language models can state falsehoods confidently and produce references that look real but are not. Before you rely on anything:
- Look up every cited book, article and DOI yourself. If you cannot find it, delete it.
- Check each statistic, date and quotation against the original source.
- Never cite a source you have not opened.
- Treat summaries of papers as a pointer, then read the paper.
An invented reference in a submission can be treated as fabrication, which is a separate and often more serious charge than plagiarism.
Common mistakes
- Assuming silence means permission. If the brief does not mention AI, ask.
- Hiding use because the rules are strict. Disclosure is usually far safer than discovery.
- Using a "humanizer" to disguise generated text. This is concealment, not editing. The AI Humanizer is built for tidying drafts you are allowed to write with AI, not for evading a policy.
- Pasting AI text and calling it a paraphrase. Rewording generated text does not make you its author. See paraphrasing vs plagiarism.
- Skipping citations for the sources behind AI claims. Cite the original sources you verified, not only the tool.
- Deleting your drafts. Version history is your best evidence of your own process.
Check your own work before submitting
A pre-submission routine takes ten minutes:
- Reread the policy for this assignment.
- List what each tool did and compare it with what is allowed.
- Verify all references and numbers.
- Run the Plagiarism Checker on your draft. It compares passages with open encyclopedic and academic sources, lists the sources it searched, and a 0% result means no overlap was found in those sources, not that the work is original.
- If the AI Content Detector is part of your review, remember it gives an estimate, not proof, and can be wrong on simple or non-native writing. If you are flagged unfairly, see falsely accused of using AI.
- Add your disclosure statement if required.
Checklist
- I know this assignment's AI rule, and I have it in writing if there was any doubt.
- Every sentence and idea in the submission is either mine or cited.
- Any generated text is permitted, marked, cited and checked.
- Every reference exists and says what I claim it says.
- My disclosure names the tool, what it did and what I did.
- My drafts and notes are saved.
Bottom line
AI use falls on a spectrum, and the dividing line is set locally. Stay on the side where you can show your own thinking and writing, disclose what tools did, verify what they produced, and when the policy is unclear, ask before you submit rather than explain afterwards.
Try the tools
Frequently asked questions
Is using ChatGPT considered plagiarism?
It can be. Submitting generated text as your own, or using it where your course bans AI, is usually treated as an integrity violation. Brainstorming or grammar help is often allowed. Your assignment brief and institutional policy decide, so check them first.
Do I have to disclose AI use for grammar fixes?
It depends on the policy. Some institutions exempt basic spelling and grammar correction, others ask you to mention any tool. If the rules are unclear, add a one-line note or ask your instructor. A short disclosure costs you nothing.
Can I list ChatGPT as an author?
Generally no. Many publishers and style guides say AI tools cannot be authors because they cannot take responsibility for the work. Disclose the tool in your methods or acknowledgements, cite it if you quote its output, and check the journal guidelines.
Will I be caught if I do not disclose AI use?
Do not rely on detection. Detectors give estimates and can be wrong in both directions, but instructors also notice style shifts, fake references and an inability to explain the work. Following the policy and keeping drafts is the safer approach.
Related guides
- What Is Plagiarism? 8 Types With Examples — Plagiarism is presenting another person's words, ideas, data or work as your own without giving credit. It can be delibe…
- How to Avoid Plagiarism: 7 Habits That Work — You avoid plagiarism by always showing where words and ideas came from. Keep a source tag on every note, put exact words…
- How Do AI Detectors Work, and How Accurate Are They? — AI detectors do not read minds or find a watermark. They measure how predictable a text is, or they use a classifier tra…
- Falsely Flagged as AI? Causes and Next Steps — If an AI detector flagged writing you did yourself, stay calm: detector scores are statistical estimates, not proof, and…
- Is Paraphrasing Plagiarism? Where the Line Is, With Four Graded Examples — Paraphrasing is not plagiarism when you restate an idea in your own words and structure and cite the source. It becomes …