SourceCited / How to cite ChatGPT & AI
Reference · Citing AI
Yes, cite AI you use. Name the model and version, give the date and prompt, and disclose that you used it. Here are the APA, MLA, and Chicago formats — and the honesty rules behind them.
If you use ChatGPT or another AI tool, cite it: name the model and version, the date, and the prompt, and disclose the use. APA treats the model as author (OpenAI as publisher); MLA templates around the prompt and tool; Chicago uses a note, like a personal communication. Above all: check your local policy, never cite AI as the source of a fact, and verify every claim against a real source.
Yes — if you use ChatGPT or another AI tool in your work, you should cite it, and the major style guides now tell you how. The core idea across all of them: treat the AI as the author/tool, name the model and version, give the date, and — because AI output is not reproducible or publicly retrievable — note that the exact response was generated for you. Below are the current formats for APA, MLA, and Chicago, plus how to handle images and the honesty rules that matter more than the punctuation.
Check your instructor's, journal's, or employer's policy first. Some prohibit AI use, some require a specific disclosure format, and that rule overrides any general style-guide convention.
APA treats the AI model as the author and the company as the publisher. The general shape:
OpenAI. (2026). ChatGPT (version) [Large language model]. https://chat.openai.com/
In text, cite it like any author-date source — for example, a parenthetical with the company name and year. APA advises that because a chatbot's reply is not retrievable by others, you should describe how you used it in your methods or text, and can include the full prompt and response in an appendix if it matters to the work.
MLA templates the citation around what the AI produced. Name the prompt as the title, the tool as the source, and include the date and URL:
"Describe the water cycle" prompt. ChatGPT, [version], OpenAI, [day month year], chat.openai.com/.
MLA is explicit that you should acknowledge functional uses of AI (editing, translating, generating) even when you would not formally cite them, and that you should vet any facts the tool gives you against reliable sources — the AI is not itself a citable authority for a fact.
Chicago generally treats AI-generated content like a personal communication: cited in a note rather than the bibliography, because the reader cannot access the original exchange. A note names the tool, the company, the date, and how it was used:
1. Text generated by ChatGPT, OpenAI, [day month year], in response to the prompt "[your prompt]."
If your discipline or publisher wants it in the bibliography, or wants the prompt disclosed, follow that — Chicago leaves room for the exchange to be shared with the reader (e.g. an appendix or a link) when reproducibility matters.
Citing an AI-generated image follows the same logic as text: name the tool and version, the date, and the prompt that produced it, and treat it as generated (not a retrievable published work). For any AI media, disclosing the exact prompt is good practice because it is the closest thing to a reproducible source. This also intersects with content provenance — C2PA credentials can record that an image was AI-generated, which supports honest disclosure.
Two very different questions share the word cite. This page is about citing ChatGPT as a source in your writing. If instead you want AI engines to cite your website, that is the rest of this site — start with getting cited by ChatGPT.
Yes, if you used it in your work — and you should disclose the use even for functional help like editing or translating. Undisclosed AI use is the real integrity issue; the exact citation format is secondary to transparency. Always check your instructor's, journal's, or employer's policy first, since it overrides general conventions.
APA treats the model as the author and the company as publisher: OpenAI. (year). ChatGPT (version) [Large language model]. URL. Cite it in text like an author-date source, and because the reply isn't retrievable by others, describe how you used it and consider including the prompt and response in an appendix.
MLA templates around what the AI produced: name the prompt as the title, then ChatGPT, the version, OpenAI, the date, and the URL. MLA also says to acknowledge functional uses of AI and to verify any facts it provides against reliable sources rather than treating the AI as the authority.
Chicago generally treats AI output like a personal communication — cited in a note, not the bibliography, because the exchange isn't publicly accessible. The note names ChatGPT, OpenAI, the date, and the prompt. If your publisher wants the prompt disclosed or a bibliography entry, follow that.
No. AI models can be confidently wrong, so never treat one as the authority for a factual claim — verify the fact against a real, citable source and cite that source instead. Cite the AI only as the tool you used, and disclose the prompt so your use is transparent.
Same logic as text: name the tool and version, the date, and the prompt that produced it, and treat it as generated rather than a retrievable published work. Disclosing the exact prompt is the closest thing to a reproducible source, and C2PA content credentials can record that the image was AI-generated.