How to disclose AI use in writing: MLA, APA and policy wording

The two most used style guides agree on the principle and differ on the mechanics.

MLA states that you should "cite a generative AI tool whenever you paraphrase, quote, or incorporate into your own work any content (whether text, image, data, or other) that was created by it" and "acknowledge all functional uses of the tool (like editing your prose or translating words) in a note, your text, or another suitable location" (MLA Style Center, how do I cite generative AI in MLA style, read 9 November 2026). MLA's August 2025 revision added that a stable, shareable chat URL is preferred and that the model name belongs in the Version element.

APA's September 2025 guidance states that "if you used AI when writing your paper, it will usually be necessary to disclose the use and cite the chat and/or AI tool", names the company as author because "AI itself cannot be an author", and says that for editing your own writing "a general reference to the tool, rather than a reference for each chat, is enough" (APA Style, citing generative AI in APA Style, Part 1, read 9 November 2026). APA's 2023 post, now archived, asked for a description of the use "in your Method section or in a comparable section of your paper" (APA Style blog, how to cite ChatGPT, read 9 November 2026).

Style guides tell you how to cite; institutions tell you whether, and what counts as misconduct when you do not. Three published policies, read on 9 November 2026, show the range.

Cambridge states that "a student using any unacknowledged content generated by artificial intelligence within a summative assessment as though it is their own work constitutes academic misconduct, unless explicitly stated otherwise in the assessment brief" (University of Cambridge, Artificial Intelligence, Education Quality and Policy Office, read 9 November 2026). Oxford states that "unauthorised use of AI falls under the plagiarism regulations and would be subject to academic penalties in summative assessments", and that where use is authorised "you should give clear acknowledgment of how it has been used in your work" (University of Oxford, use of generative AI tools to support learning, read 9 November 2026). Harvard's Office of Undergraduate Education offers instructors three sample policies and states in the permissive ones that "any such use must be appropriately acknowledged and cited" and that "you bear the final responsibility" for AI output (Harvard Office of Undergraduate Education, AI guidance and FAQs, read 9 November 2026).

The word that recurs is "unacknowledged". None of the three bans acknowledged use outright; all three make the failure to acknowledge the offence. Whether a given use is permitted is a separate question, answered on the page about whether using an AI humanizer is cheating.

A disclosure has four parts: the tool and model, what it did, where in the work, and what remained yours. Keep it to two or three sentences and put it where the policy or style guide says: a note for MLA, the Method section or author note for APA, the acknowledgements or a cover sheet where the course specifies one.

The last row is the one most readers of this site are asking about. A humanizer is a functional use under MLA's definition and an editing use under APA's, so both guides require you to say so. If saying so would be a problem under your course policy, the policy is telling you not to use the tool, and no wording fixes that. The academic writing page and the students page start from that rule.

MLA's published example for a paraphrased response reads, in its structure: the prompt in quotation marks followed by "prompt", then the tool name in italics, the model, the company, the date generated, and the shareable chat address, for example a prompt about a novel's themes, ChatGPT, model GPT-4o, OpenAI, 23 Sept. 2024, with a chatgpt.com share link (MLA Style Center, updated and revised post, 13 August 2025, read 9 November 2026). The in-text citation uses a short form of the prompt in parentheses.

APA's published example for a specific chat reads: the company, the full date of the chat in parentheses, the chat title in italics followed by a bracketed description such as "[Generative AI chat]", the tool or model name, and the chat URL; the in-text citation is the company and year, for example (OpenAI, 2025). For a tool cited generally, APA drops the date to the year and the title to the tool name with "[Large language model]" after it, and no longer asks for a version number, read 9 November 2026.

Two cautions from the guides themselves. MLA warns that AI tools "can also hallucinate, or make up, sources", so a source the tool named must be read and cited directly, never through the tool. APA says prompts belong in the text or an appendix, not the reference, because a prompt "doesn't help readers retrieve the work". Disclosure is about letting a reader see what you did; a disclosure that hides the prompt or launders a fabricated source fails that test.

A disclosure statement is evidence of honesty, not a licence. The guides hub holds the related pages, including what to do if a detector flags writing that is yours.