AI-Assisted Essay Writing: How to Format Citations for ChatGPT and AI Sources in 2026

The problem with AI-generated text

ChatGPT, Gemini, and Copilot are changing how people write essays. These tools generate text on almost any topic, helping with everything from the first brainstorm to the final draft. This creates a problem for academic standards: nobody is quite sure how to cite a machine.

The core issue revolves around plagiarism, academic integrity, and acknowledging assistance. Simply submitting AI-generated text as your own work is, undeniably, plagiarism. But even when used as a tool for idea generation or drafting, failing to cite AI obscures the extent of its contribution. It's a messy situation because the guidelines are still evolving, and universities are scrambling to catch up.

AI detection tools are getting better, even if they still make mistakes. Getting flagged is a real risk. Being honest about using AI is better than crossing your fingers and hoping a detector misses your work. If you cite the tool, you aren't hiding anything.

How Purdue and UMD handle AI

Currently, two leading university writing centers – Purdue University and the University of Maryland – offer guidance on citing AI-generated content. Both institutions recommend treating AI, specifically tools like ChatGPT, as an author. This might seem counterintuitive, as AI isn’t a person, but it reflects the significant role these tools play in the creation of the text.

Purdue’s Online Writing Lab (OWL) says to list 'ChatGPT' as the author and 'OpenAI' as the publisher. The University of Maryland follows the same logic. OpenAI hasn't said they want to be listed as authors, so the citation feels a bit clunky, but it is the standard for now.

The reasoning behind this approach is that AI isn't simply a spellchecker or grammar tool. It actively generates content, making a demonstrable contribution to the final product. This contribution needs to be acknowledged. Both Purdue and UMD stress the importance of providing enough information to allow readers to locate the AI model used and, if possible, the specific prompts that generated the text.

The core principle, according to both guides, is to be as transparent as possible about your use of AI. This isn't about avoiding trouble with detection software; it's about upholding academic honesty and allowing your work to be evaluated fairly. It also establishes a clear record of the creative process, acknowledging where your ideas originated and where AI assisted.

Treating AI as an author in MLA 9

Let’s get specific. Here’s how to format citations for ChatGPT and similar AI tools using the MLA 9th edition. Remember, the key is to treat the AI as an author. For an in-text citation, you’ll typically include the AI's name (ChatGPT) and, if relevant, a timestamp indicating when the response was generated. For example: (ChatGPT, Feb 29, 2024).

The corresponding Works Cited entry would look like this: ChatGPT. OpenAI, 2023. This is where it gets a little strange listing a company as the publisher for an AI model. If you’ve accessed ChatGPT through a specific platform or interface, you can include that information as well. For example, if using ChatGPT via Microsoft Copilot, you could add 'Accessed through Microsoft Copilot'.

If you’re directly quoting ChatGPT, include the prompt you used in the citation. This provides context and allows readers to understand how the AI generated the response. Example in-text citation: (ChatGPT, responding to prompt “Explain the concept of quantum entanglement,” Feb 29, 2024).

For paraphrased content, you still need to cite ChatGPT, but you don’t need quotation marks. The in-text citation would still include the AI's name and date. For AI-assisted idea generation, where the AI helped you brainstorm but didn’t directly produce text, you can acknowledge its contribution in a note or in your methodology section, stating that you used ChatGPT to explore different perspectives. The level of detail will depend on the extent of AI assistance. A simple 'Brainstorming assistance provided by ChatGPT' might suffice for minor contributions.

1. Direct quotes: (ChatGPT, responding to prompt “Explain the concept of quantum entanglement,” Feb 29, 2024)
2. Works Cited: ChatGPT. OpenAI, 2023.
3. In-text Citation (Paraphrase): (ChatGPT, Feb 29, 2024)
4. AI-Assisted Idea Generation: Acknowledge in a note or methodology section.

APA 7: AI as a research tool

APA 7th edition takes a different tack. Instead of treating AI as an author, it recommends treating it as a tool. This reflects a more pragmatic view of AI’s role in research. The emphasis in APA is on describing the tool and the prompts used to generate the content in the method section of your paper.

Instead of an author entry, you’d describe ChatGPT in the text. For example: “Text was generated using ChatGPT (OpenAI) with the following prompt: 'Summarize the key findings of the study by Smith et al. (2023)'”. The date of access isn't typically included, as the AI model itself is constantly evolving.

Crucially, APA emphasizes the importance of transparency. You must clearly state how you used the AI, what prompts you provided, and what modifications, if any, you made to the generated text. This allows readers to assess the validity and reliability of your work.

APA doesn’t have a formal citation for the reference list yet. Instead, you describe the tool in your methodology. This treats AI like a piece of lab equipment or statistical software. It is a practical way to handle it, provided you describe exactly what the tool did.

Other citation styles

Citation styles like Chicago, Harvard, and IEEE are still grappling with how to address AI-generated content. Guidance is often lacking or inconsistent. Many style guides are currently recommending adapting the principles of MLA or APA – either treating AI as an author or as a tool – depending on the extent of its contribution.

Because the landscape is so fluid, the most important advice is to check with your instructor or the specific requirements of the publication you’re submitting to. They may have specific guidelines for citing AI that supersede general style guide recommendations. It's always better to err on the side of caution and seek clarification if you're unsure.

Don’t assume that a particular style guide has a definitive answer. The situation is evolving rapidly, and new guidance is likely to emerge in the coming months and years. Staying informed and being adaptable is essential.

Citing your prompts

This is where things get really interesting – and complex. What about the prompts themselves? If you’ve spent significant time crafting and refining a prompt to achieve a specific result, can that prompt be considered a form of intellectual contribution? I believe it can, especially for iterative prompts that demonstrate a clear thought process.

In MLA, you can cite the prompt as a “personal communication,” including it in the Works Cited as “Prompt to ChatGPT, date.” In APA, you would describe the prompt in detail within the method section of your paper, treating it as part of the research procedure. The key is to demonstrate the effort and thought that went into formulating the prompt.

For complex prompts that involve multiple revisions and refinements, consider including a full transcript of the prompt engineering process in an appendix. This provides a complete record of your interaction with the AI and allows readers to understand how the generated text was produced. This is particularly important if the prompt significantly influenced the outcome.

Acknowledging the prompts isn't just about avoiding plagiarism; it's about recognizing the skill and creativity involved in prompt engineering. It's a new form of literacy, and it deserves to be acknowledged.