AI and learning

How to cite ChatGPT in APA: a step-by-step guide

The APA reference for ChatGPT with author, date, version, descriptor, and URL, plus the in-text citation, prompt records, and frequent mistakes.

The short answer: what the reference must identify

To cite ChatGPT in APA using the linked official article, identify OpenAI, the version year, the model name, the version in parentheses, the descriptor [Large language model], and the URL. Add an in-text citation and explain the prompt you used and the response you included. Generated text is treated as an algorithm’s output, not a personal communication from a human author.

The cited page is archived guidance and points to an update published in September 2025. This guide explains its structure with a generic template. Before submitting, check current official guidance and your institution’s instructions. A correct reference does not authorize AI use that your course prohibits.

Six pieces to check

Organize your check around six pieces: author, year, title, version, descriptor, and URL. The version and descriptor accompany the model title; checking them separately helps you catch omissions.

Do not fill in details from memory. Identify the tool you used and retain version information you can verify. If a detail is unavailable, check how the applicable instructions handle it instead of inventing a date or copying an older version.

  • Author: OpenAI, following the reference explained in the APA Style article.
  • Year: the year of the version used, not automatically the year you submit the paper.
  • Title: the model name, italicized in the final reference.
  • Version: in parentheses, using verifiable identification and the required format.
  • Descriptor: [Large language model], following the cited article’s structure.
  • URL: the relevant access address. Do not repeat the publisher when it matches the author.

A generic template without an invented version

The following line is a template, not a reference ready to submit. Replace the bracketed year, version, and URL placeholders with verified details. The descriptor [Large language model] is part of the reference structure:

OpenAI. ([version year]). ChatGPT ([version used]) [Large language model]. [access URL]

For example, when preparing a submission from a saved conversation, first check which version details are available. Do not copy the year from APA’s historical example for a different interaction. Then review these pieces before adding the entry to your final reference list:

  • The author should identify the model’s creator, not someone who wrote a tutorial about it.
  • The in-text citation year should match the reference entry.
  • The version label should describe what you used without guessing a release.
  • Italicize the model name and retain the descriptor’s square brackets.
  • Check the URL and current guidance before submitting the reference.

In-text citation and where to describe your use

A parenthetical citation takes the form (OpenAI, [year]); a narrative citation uses OpenAI ([year]) within the sentence. You can combine these options as appropriate. Replace the placeholder with the year matching the reference. Do not imply that OpenAI researched or endorsed a claim merely because ChatGPT generated it.

The APA Style article suggests describing research use of the tool in the Method section or a comparable section. For essays and reviews, it notes that the introduction may be appropriate. Include the prompt and the relevant response, while also following the assignment’s specific instructions.

Document the prompt and the generated text

Keep the exact text you used. Repeating a prompt may produce a different response, so a new query does not necessarily reconstruct the earlier one. For a long response, the article discusses an appendix or supplemental material, mentioned in the body of the paper.

A useful record includes the date, tool, known version, prompt, relevant response, and later changes. Also keep the reference for the document used to verify the content. For example: “I requested a comparison of two definitions, noticed a missing condition, and corrected it using the course chapter.” This explains your intervention; it does not turn the AI response into scientific evidence.

Common mistakes when citing ChatGPT

Check citation formatting separately from content quality. A reference can be correctly formatted while the cited response contains mistakes.

If ChatGPT suggests a source, open the original and check its author, title, and content. A working link is not enough; the source must support the specific claim. To support an academic argument, consult and cite the material you actually read, without attributing the chatbot’s unsupported statements to it. Also check for these formatting and documentation errors:

  • Inventing a date or version to make an entry look complete.
  • Using the prompt as the model’s title in the reference.
  • Omitting the descriptor or unnecessarily repeating the publisher’s name.
  • Describing generated text as a personal communication from a human author.
  • Failing to distinguish a saved response from a new response to another prompt.
  • Assuming a saved response cannot be cited again: you can refer back to it where appropriate.

When citing is not enough

If the course prohibits AI for the activity, adding a reference does not make that use acceptable. If only certain assistance is allowed, stay within that scope. Permission to practice questions does not necessarily authorize generating paragraphs for an essay.

APA guidance and institutional policies can change. The linked page states that it is archived and directs readers to later guidance. Use it to understand the example explained here, then check current instructions before applying the structure to an actual submission.

Before submitting, check the following:

  • The course policy: read the syllabus, the brief, or the rubric for what it says about generative AI and under which conditions.
  • The date of the guidance: the page you read today may have been revised, so keep the date you consulted it.
  • The required style: apply the guide your institution specifies for this submission.
  • The evidence you kept: prompt, generated fragment, model version, and a note of what you corrected.
  • The sources the model gave you: open them and confirm that they say what they are credited with before you rely on them.

Keep references with your study notes

Lumia AI lets you create or edit flashcards, generate AI cards and quizzes from notes or PDFs, organize them into books, and review with spaced repetition based on your ratings. It also offers a chat tutor, learning paths, Anki .apkg deck import, and web and mobile apps.

For this workflow, prepare review material and check each answer against its source. Keep references and your AI-use record in your own files; do not assume a flashcard automatically contains a valid citation. Before including an idea in an assignment, return to the original and check both its content and how to acknowledge the assistance you received.

Sources and scope of this guide

These references provide context on studying and learning. Examples are illustrative; they are not an evaluation of Lumia AI and do not guarantee grades or retention rates.

Editorial approach: the body names the source beside evidence-based claims; this list identifies the document, link, and scope consulted. We distinguish those references from our own examples. If you find an error, include the page and sentence when you contact the team.

Keep learning

AI news · 5 min read

PISA 2025 and AI: what the science results mean

The OECD published PISA 2025 with data from more than 760,000 students. Explore the AI-use comparisons, their limits, and a practical study routine.

Lumia AI teamPublished:

Prepare flashcards and keep your references

Generate flashcards and quizzes from your notes in Lumia AI, check the content, and keep assignment references in your own files.

Get started free