Blog/May 25, 2026·9 min read

How to Cite ChatGPT in Academic Writing: APA, MLA, Chicago (2026)

By
Writer & Editor · Updated May 25, 2026

Quick Answer

Cite ChatGPT in academic writing using your style guide's format for software or personal communications. APA 7th: OpenAI. (2024). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com. MLA 9th treats it as software with the prompt as the title. Chicago treats it as personal communication. Always disclose AI use in your methodology section.

Universities, journals, and publishers updated their AI policies in 2023 and 2024. Most now require disclosure when ChatGPT, Claude, or Gemini contributed to a paper. This guide covers official APA 7, MLA 9, Chicago 17, and IEEE formats with full examples, plus the bigger problem of hallucinated citations.

Why You Must Cite AI in Academic Work

  • Academic integrity. Submitting AI text without disclosure violates the integrity codes of nearly every accredited university.
  • Institutional policy. APA, MLA, Chicago, and most universities now have specific guidance. Ignoring it is on you.
  • Reproducibility. AI output changes by version and date. Citing the version lets others verify your claims.
  • Plagiarism risk. Even though AI is not a human author, treating its output as your own writing can be classified as fabrication or contract cheating, both serious offenses.
  • Professional protection. Disclosed AI use is rarely punished. Undisclosed AI use is.

If you used ChatGPT for your essay, you need to cite it. Even if you used it only for brainstorming. Even if you rewrote every sentence. Disclosure protects your record. Before submitting, you can also run your draft through our free AI Detector to see what a reviewer might see.

APA 7th Edition Format

The American Psychological Association released official ChatGPT citation guidance in April 2023 and updated it through 2024. APA treats large language models as software.

APA Reference Format

Author. (Year). Name of tool (Version) [Type of model]. URL

APA In-Text Citation

(OpenAI, 2024)
OpenAI (2024) generated...

APA Example: ChatGPT

OpenAI. (2024). ChatGPT (Mar 14 version) [Large language model].
   https://chat.openai.com

APA Example: Claude

Anthropic. (2024). Claude 3.5 Sonnet (June 2024 version) [Large language model].
   https://claude.ai

APA Example: Gemini

Google. (2024). Gemini 1.5 Pro [Large language model].
   https://gemini.google.com

The APA Style Blog (2023) clarified that ChatGPT prompts and outputs are non-recoverable, so the reference cites the tool itself. Include the actual prompt text and key outputs in an appendix or methodology section.

MLA 9th Edition Format

The MLA Style Center released AI citation guidance in March 2023. MLA 9 treats AI as software, with the prompt as the title.

MLA Works Cited Format

"Prompt text in quotes" prompt. Tool name, Version, Publisher, Date, URL.

MLA In-Text Citation

("First few words of prompt")

MLA Examples for Each Tool

ChatGPT:

"Describe the symbolism of the green light in The Great Gatsby" prompt.
   ChatGPT, 14 Mar. version, OpenAI, 14 Mar. 2024, chat.openai.com.

Claude:

"Summarize the key arguments in Kant's Categorical Imperative" prompt.
   Claude 3.5 Sonnet, Anthropic, 10 June 2024, claude.ai.

Gemini:

"Compare French and Russian Revolutions causes" prompt. Gemini 1.5 Pro,
   Google, 22 Apr. 2024, gemini.google.com.

Chicago / Turabian Format

The Chicago Manual of Style (17th edition) and the online supplement treat AI as either a personal communication (notes-bibliography style) or as software (author-date style).

Chicago Notes-Bibliography Format

Footnote:
1. Text generated by ChatGPT, OpenAI, March 14, 2024,
   https://chat.openai.com.

Bibliography (optional, often omitted for personal communications):
OpenAI. ChatGPT. Large language model. March 14, 2024.
   https://chat.openai.com.

Chicago Author-Date Format

In-text:
(OpenAI 2024)

Reference list:
OpenAI. 2024. ChatGPT (March 14 version). Large language model.
   https://chat.openai.com.

Chicago Examples

Footnote with prompt context:

2. ChatGPT, response to "Explain Bayesian inference for an undergraduate
   statistics course," OpenAI, April 22, 2024, https://chat.openai.com.

Claude in Chicago notes:

3. Text generated by Claude 3.5 Sonnet, Anthropic, June 10, 2024,
   https://claude.ai.

Chicago does not require AI sources in the bibliography when treated as personal communication. Many advisors prefer them in the reference list anyway for transparency.

IEEE Format (for STEM and CS papers)

IEEE updated its citation guidance to cover generative AI in 2023. Use the software citation template.

Reference:
[1] OpenAI, "ChatGPT (Mar 14 version)," Large language model, 2024.
    [Online]. Available: https://chat.openai.com

In-text:
The model output [1] indicates...

How to Disclose AI Use in Methodology

Citing the tool is not enough. You also need a disclosure paragraph in your methodology or acknowledgments. Here is a sample:

We used ChatGPT (GPT-4o, May 2024 version, OpenAI) to brainstorm research questions and draft the literature review introduction. All prompts and the conversation history are available in Appendix A. The first author reviewed, fact-checked, and rewrote all AI-generated text. Final responsibility for accuracy and argument rests with the authors. No AI tools generated data, analyzed results, or wrote the discussion section.

What to include:

  • Tool name, version, publisher, and access date
  • Which sections used AI assistance
  • Which prompts (or a representative sample in an appendix)
  • What was kept, edited, or discarded
  • Who reviewed and fact-checked the output
  • What AI did not do

Where in your paper:

  • Methodology section (always)
  • Acknowledgments (most journals require it here too)
  • Appendix with full prompt history (recommended)
  • Submission checkbox or AI disclosure form (where required)

Common Mistakes When Citing ChatGPT

  • Treating output as a primary source. ChatGPT's output is generated text, not a primary source. Cite the underlying paper, book, or study, not the AI summary.
  • Citing ChatGPT as a person or author. AI is not an author under any major style guide. List it as software.
  • Forgetting the date and version. AI output changes weekly. Without a date and version, your citation is unreproducible.
  • Not verifying AI-cited references. ChatGPT, Claude, and Gemini all hallucinate sources. Verify every citation before using it.
  • Hiding AI use. The biggest mistake. Most academic misconduct cases involving AI start with undisclosed use, not the use itself.
  • Citing a screenshot URL. A shared conversation URL is fine, but a screenshot is not citable on its own. Use the tool's official domain.

For self-checking your draft, run it through our free AI Detector before submitting. The detector runs in your browser and helps you see what a reviewer might see.

Hallucinated Citations: The Bigger Problem

ChatGPT, Claude, and Gemini all fabricate references. This is not a rare bug. It is a regular failure mode driven by how language models work: they predict plausible text, and a plausible-looking citation is just as easy to generate as a real one.

Common hallucination patterns:

  • Real authors paired with fake paper titles
  • Real journals paired with fake DOIs
  • Real DOIs that resolve to a different paper
  • Plausible-sounding citations that do not exist at all
  • Page numbers that do not match the actual book

How to verify every citation:

  1. Search Google Scholar for the exact title.
  2. Click through the DOI to confirm it resolves to the cited paper.
  3. Check the publication year against the journal's archive.
  4. For books, verify the page number in a physical or library copy.
  5. For quotes, find the original in the actual source. If you cannot, do not use the quote.

As part of your review process, consider how the AI output reads overall. Our guide on how to humanize AI text covers the editorial signals that also expose generic citations and weak evidence.

School Policy Quick Reference

Policies vary widely across institutions. Three broad categories:

  • Disclose and allow: Most universities since 2023. AI use is permitted if disclosed in methodology and acknowledgments. Stanford, MIT, Harvard, and most R1 institutions fall here.
  • Ban with rubric: Some courses prohibit AI use entirely. Common in writing-intensive courses, freshman composition, and high school AP courses.
  • Draft-only: AI permitted for outlining, brainstorming, and editing, but not for generating final text. Increasingly common in undergraduate humanities.

Always check the syllabus, the course honor code, and your institution's academic integrity policy. When in doubt, ask your instructor in writing and save the response. For teachers, our teacher guide to detecting AI essays covers the practical detection workflow.

AI in academic writing is here to stay. Citing it properly, disclosing your use, and verifying every reference are the three habits that keep your work credible and your transcript clean. Pair that with our guide on humanizing AI text and the free AI Detector for a complete pre-submission workflow.

Sources

  1. American Psychological Association (2023). How to cite ChatGPT. APA Style Blog.
  2. Modern Language Association (2024). How do I cite generative AI in MLA style? MLA Style Center.
  3. University of Chicago Press (2017). The Chicago Manual of Style (17th ed.) and Online Q&A on AI Citation.
  4. International Center for Academic Integrity (2024). Best Practices for Academic Integrity with Generative AI.
  5. Stanford HAI (2024). Academic Integrity in the Age of Generative AI. Stanford Institute for Human-Centered AI.
  6. Pew Research Center (2024). About 1 in 5 U.S. teens who have heard of ChatGPT have used it for schoolwork.

Check your word count and readability with our free Word Counter.

Open Word Counter

Related Guides

AI Writing

How to Humanize AI Text

9 research-backed techniques to make ChatGPT, Claude, or Gemini text sound natural.

Education

Teacher's Guide to Detecting AI Essays

A practical 5-step workflow for teachers to identify AI-written essays in 2026.

Academic

Abstract Word Count Guide

Word count rules for academic abstracts across journals, dissertations, and conference papers.

Frequently Asked Questions

If your institution or publisher requires it, yes. Most universities updated their academic integrity policies in 2023 and 2024 to require disclosure of AI use. Even where it is not required, citing protects you from plagiarism accusations and supports reproducibility. When in doubt, cite and disclose.

APA treats ChatGPT as software. The reference is: OpenAI. (2024). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com. In-text: (OpenAI, 2024). Include the date you accessed it and the version if you can determine it. Describe your prompts in the methodology section.

MLA 9 treats ChatGPT as software with a description of the prompt. Works Cited: 'Describe the symbolism in The Great Gatsby' prompt. ChatGPT, 14 Mar. version, OpenAI, 14 Mar. 2024, chat.openai.com. In-text: ('Describe the symbolism'). Include the prompt itself as the title.

Chicago (17th edition) treats AI as a personal communication or unpublished source. Notes-Bibliography: 1. Text generated by ChatGPT, OpenAI, March 14, 2024, https://chat.openai.com. Author-Date: (OpenAI 2024). Include the model version and access date. Chicago does not require AI sources in the bibliography for personal communications.

No. ChatGPT is not an author. Authorship requires accountability for the content, and a language model cannot be held accountable. All major style guides treat AI as a tool or software, not as an author. Listing ChatGPT as a co-author is rejected by APA, MLA, Chicago, Nature, Science, and most journals.

You need: the tool name (ChatGPT), the company (OpenAI), the model version (GPT-4o, Mar 14 version, etc.), the date you accessed it, the prompt you used (or a description), and the URL (chat.openai.com). Save a screenshot or export of the conversation for your records.

Use the same format as ChatGPT, swapping the tool name and publisher. For Claude: Anthropic. (2024). Claude 3.5 Sonnet [Large language model]. https://claude.ai. For Gemini: Google. (2024). Gemini 1.5 Pro [Large language model]. https://gemini.google.com. Apply your style guide's rules for software.

Two places. First, in the methodology section, describe which AI tools you used, which prompts, and what you kept or changed. Second, in the acknowledgments or a dedicated AI disclosure statement, name the tools and the scope of use. Some journals also require a checkbox on submission.

Often no. ChatGPT, Claude, and Gemini all fabricate citations regularly, including DOIs, page numbers, journal titles, and even author names. Verify every reference through Google Scholar or your library database before including it. Hallucinated citations are one of the most common ways academic misconduct cases start.

Then do not use it for graded work, or use it only as your policy allows (often brainstorming and outlining only). Read your specific course syllabus and academic integrity policy. Penalties for undisclosed AI use range from a zero on the assignment to expulsion. When unsure, ask your instructor in writing.

Most do with disclosure. Nature, Science, JAMA, Elsevier, and Wiley all permit AI assistance for editing or drafting if disclosed in the methods or acknowledgments. None permit AI as an author. Specific journals may have stricter rules. Always check the submission guidelines for your target venue.

Yes. Save the full conversation as a PDF, screenshot, or shared link (ChatGPT and Claude both support shareable conversation URLs). Some style guides recommend including key conversations in an appendix. This protects you against later questions about what the AI generated versus what you wrote.