The MLA’s method for citing sources uses a template of core elements—standardized criteria that writers can use to evaluate sources and create works-cited-list entries based on that evaluation. That new technologies like ChatGPT emerge is a key reason why the MLA has adopted this approach to citation—to give writers flexibility to apply the style when they encounter new types of sources. In what follows, we offer recommendations for citing generative AI, defined as a tool that “can analyze or summarize content from a huge set of information, including web pages, books and other writing available on the internet, and use that data to create original new content” (Weed).
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
- acknowledge all functional uses of the tool (like editing your prose or translating words) in a note, your text, or another suitable location
- take care to vet the secondary sources it cites (see example 5 below for more details)
See below for specific examples. And keep in mind: the MLA template of core elements is meant to provide flexibility in citation. So if you find a rationale to modify these recommendations in your own citations, we encourage you to do so. We’ve opened this post up for commenting, so let us know what you think and how you’re using and citing generative AI tools!
Using the MLA Template
Author
We do not recommend treating the AI tool as an author. This recommendation follows the policies developed by various publishers, including the MLA’s journal PMLA.
Title of Source
Describe what was generated by the AI tool. This may involve including information about the prompt in the Title of Source element if you have not done so in the text.
Title of Container
Use the Title of Container element to name the AI tool (e.g., ChatGPT).
Version
Name the version of the AI tool as specifically as possible. For example, the examples in this post were developed using ChatGPT 3.5, which assigns a specific date to the version, so the Version element shows this version date.
Publisher
Name the company that made the tool.
Date
Give the date the content was generated.
Location
Give the general URL for the tool.1
Example 1: Paraphrasing Text
Passage in Source
Paraphrased in Your Prose
While the green light in The Great Gatsby might be said to chiefly symbolize four main things: optimism, the unattainability of the American dream, greed, and covetousness (“Describe the symbolism”), arguably the most important—the one that ties all four themes together—is greed.
Works-Cited-List Entry
“Describe the symbolism of the green light in the book The Great Gatsby by F. Scott Fitzgerald” prompt. ChatGPT, 13 Feb. version, OpenAI, 8 Mar. 2023, chat.openai.com/chat.
Example 2: Quoting Text
Passage in Source
Quoted in Your Prose
When asked to describe the symbolism of the green light in The Great Gatsby, ChatGPT provided a summary about optimism, the unattainability of the American dream, greed, and covetousness. However, when further prompted to cite the source on which that summary was based, it noted that it lacked “the ability to conduct research or cite sources independently” but that it could “provide a list of scholarly sources related to the symbolism of the green light in The Great Gatsby” (“In 200 words”).
Works-Cited-List Entry
“In 200 words, describe the symbolism of the green light in The Great Gatsby” follow-up prompt to list sources. ChatGPT, 13 Feb. version, OpenAI, 9 Mar. 2023, chat.openai.com/chat.
While we’ve provided fairly detailed descriptions of the prompts above, a more general one (e.g., Symbolism of the green light in The Great Gatsby prompt) could be used, since you are describing something that mimics a conversation, which could have various prompts along the way.
Example 3: Citing Creative Visual Works
If you are incorporating an AI-generated image in your work, you will likely need to create a caption for it following the guidelines in section 1.7 of the MLA Handbook. Use a description of the prompt, followed by the AI tool, version, and date created:
Fig. 1. “Pointillist painting of a sheep in a sunny field of blue flowers” prompt, DALL-E, version 2, OpenAI, 8 Mar. 2023, labs.openai.com/.
You can use this same information if you choose to create a works-cited-list entry instead of including the full citation in the caption (see MLA Handbook, sec. 1.7).
Example 4: Quoting Creative Textual Works
If you ask a generative AI tool to create a work, like a poem, how you cite it will depend on whether you assign a title to it. Let’s say, for example, you ask ChatGPT to write a villanelle titled “The Sunflower” that—you guessed it!—describes a sunflower and then quote it in your text. Your works-cited-list entry might look like this:
“The Sunflower” villanelle about a sunflower. ChatGPT, 13 Feb. version, OpenAI, 8 Mar. 2023, chat.openai.com/chat.
If you did not title the work, incorporate part of or all of the first line into the description of the work in the Title of Source element:
“Upon the shore . . .” Shakespearean sonnet about seeing the ocean. ChatGPT, 13 Feb. version, OpenAI, 8 Mar. 2023, chat.openai.com/chat.
For guidance on using descriptions and text from the work itself in the Title of Source element, see the MLA Handbook, 5.28 and 5.29.
Example 5: Citing Secondary Sources Used by an AI Tool
You should also take care to vet the secondary sources cited by a generative AI tool—with the caveat that AI tools do not always cite sources or, when they do, do not always indicate precisely what a given source has contributed. If you cite an AI summary that includes sources and do not go on to consult those sources yourself, we recommend that you acknowledge secondary sources in your work.
For example, let’s say that you ask Bing AI to explain the concept of the political unconscious, citing sources, and it provides the following answer:
Let’s say that you then decide to quote from the final sentence. You need to click through to the source listed in the note in order to get more information than just a URL for the source. There, you will read the following:
Now, you can treat Oxford Reference as your source since Bing AI was merely a research conduit to the source (see MLA Handbook 5.34 for more information). If for some reason you want to treat a source cited in a generative AI tool as an indirect source–and you know it is, in fact, the source for the information provided by the AI, follow the guidance in section 6.77 of the MLA Handbook.
Note
1. At the time of writing this post, ChatGPT doesn’t have a built-in feature to create a unique URL to the conversation. However, an outside tool like the Chrome extension ShareGPT can generate such a link. If you use that type of outside tool, include the unique URL that the tool generates instead of the general URL.
DALL-E allows users to download the AI-generated images they create or generate a publicly-available URL that leads to an image. If you choose to create a shareable link for an image you generate with DALL-E (or other similar AI image generators), include that unique URL that leads to the image instead of the general URL.
Works Cited
MLA Handbook. 9th ed., Modern Language Association of America, 2021. MLA Handbook Plus, 2021, mlahandbookplus.org/.
Weed, Julie. “Can ChatGPT Plan Your Vacation?” The New York Times, 16 Mar. 2023, www.nytimes.com/2023/03/16/travel/chatgpt-artificial-intelligence-travel-vacation.html.
44 Comments
Caitlin Ratcliffe 21 March 2023 AT 06:03 PM
Thank you for this guidance about ChatGPT! This is very timely and helpful. Can you please confirm the order of the core elements in your guidance here? Specifically the "Publisher" element and the "Date" element. Thank you!
Laura Kiernan 22 March 2023 AT 12:03 PM
Thank you for your question about the order of the Publisher and Date elements. We have corrected this post to show that the Publisher element should be before the Date element in works-cited-list entries.
Christiana Salah 22 March 2023 AT 06:03 PM
While you say to vet the sources, which broadly covers this, it might be helpful to explicitly state in your guide that ChatGPT will invent plausible-looking sources. These fake citations use the names of real publication venues and sometimes the names of real scholars in the correct field.
For example, I tested ChatGPT by asking for a paragraph on a recent novel citing a peer-reviewed source. It created a plausible citation for an essay in a paywalled new issue of a relevant journal. Only by investigating the author's name and finding them not to exist was I able to tell the citation was invented. It appears ChatGPT will cite real sources when it finds such, but invent one if it does not.
Diana Tsang 22 March 2023 AT 08:03 PM
Thank you for this timely reference!
Could you kindly confirm if the format of the Version Date (Feb. 13 version) is different from the MLA's usual format of Date (not "13 Feb." version) on purpose?
Laura Kiernan 23 March 2023 AT 11:03 AM
Thank you for pointing out that error. We have corrected this post so the date in the Version element follows the day-month-year date style: 13 Feb. version.
Diana Tsang 22 March 2023 AT 08:03 PM
As the Title of Container (e.g. ChatGPT) is not a just description. Shall it be italicized in the works cited entry?
Laura Kiernan 23 March 2023 AT 11:03 AM
Thank you for pointing out that error. ChatGPT should be italicized because it is the name of a piece of software, so we have corrected this post to reflect that.
Tarn 28 March 2023 AT 09:03 PM
Should DALLE-E be italicised in the above example too?
Fig. 1. “Pointillist painting of a sheep in a sunny field of blue flowers” prompt, DALLE-E, version 2, OpenAI, 8 Mar. 2023, labs.openai.com/.
Laura Kiernan 29 March 2023 AT 09:03 AM
Thank you for alerting us to that error. Yes, that AI tool name should be italicized, so we have corrected the post to reflect that. We have also corrected the spelling to be DALL-E.
Carine Ndzewiyi 03 January 2024 AT 12:01 AM
Thank you for being my eye opener. Yes DALL- E should be italicized
Amy Curtis 24 March 2023 AT 12:03 PM
How would you suggest to cite ChatGPT when using it to summarize personally generated data sets? For example in lab experiments, or observational data.
Christian Schmidt 03 April 2023 AT 04:04 PM
Thank you for these guidelines. Question on Example 3 "Citing Creative Visual Works". You are NOT suggesting to include the link to a retrievable/published version of such an image, instead you advice to provide the general URL of the tool (here: labs.openai.com/). But: You are currently the only Style provider who has touched on the retrievability issue in a bit more differentiated way (in your note to Example 5). You are even mentioning workarounds like ShareGPT for text there, which is appreciated. But I don't understand, why you would not suggest the same for images, since some of the image creation tools even allow for publishing them - no workaround needed. OpenAI's DALL-E does it (example: https://labs.openai.com/s/jrbn6Dl8aRTDvS4Kobny33n2), and others like Midjourney allow for that, too. Any reason why that is not included here?
Laura Kiernan 04 April 2023 AT 10:04 AM
Thank you for bringing that to our attention. We have updated the note in the post to acknowledge that if someone chooses to create a shareable link for an image generated with DALL-E (or other similar AI image generators), that unique URL should be included instead of the general one.
Anna Mills 11 April 2023 AT 05:04 PM
Here's a suggestion for a way to foreground the human critics who are the real sources of the ideas about the symbolism of the green light in the above Fitzgerald example:
Unknown human authors statistically remixed by ChatGPT, 13 Feb. version, OpenAI, 8 Mar. 2023, chat.openai.com/chat. “Describe the symbolism of the green light in the book The Great Gatsby by F. Scott Fitzgerald” prompt.
Jennifer 03 October 2024 AT 06:10 AM
... except that 'statistically remixing' existing works is not what LLMs do. Probabilities are involved, so perhaps 'statistically' comes close, but the only thing being 'remixed' is the whole of human language, as represented by the text of the entire Web ...
Thomas Basbøll 12 April 2023 AT 04:04 AM
Christiana's point above deserves an answer. In example 2, you provide a screenshot of such "hallucinated" references, but you suggest quoting GPT uncritically describing it as a "list of scholarly sources". Is this good practice?
E A 15 June 2023 AT 02:06 AM
Remind your students that GPT is not a scholarly source, nor a source that's any better (actually worse) than wikipedia. It's a generative pre-trained transformer. I can't find anything in the AI literature or specs that is directing it to use actual, scholarly, or peer-reviewed, or even actual sources, versus "faking" them, as we've seen in the comments, above. AI takes no responsibility for this, nor do, it seems, its "handlers" (originators, developers). This is parallel to when Zuckerberg, et. al, took no responsibility for FB postings pushing disinformation (esp re elections) , Russians trolls, or credibility. None, zero, zip, whatsoever.
AI also cannot be "accused" of plagiarism or dishonesty, nor does it need to experience any consequences, unlike humans.
It is responsible for nothing, and to no one. Humans, however, are.
Davis Oldham 17 April 2023 AT 10:04 AM
I agree with Thomas and Christiana. The first source cited in Example 2, Bruccoli, does not appear to exist, from what I can see in the table of contents for the cited container:
https://worldcat.org/title/10020427
Bruccoli is an actual Fitzgerald scholar but the cited title does not appear in the cited collection of critical essays.
I think it's imperative that a resource on citing chatGPT should address this.
E A 15 June 2023 AT 02:06 AM
GPT is not a scholarly source, nor a source that's any better (actually worse) than wikipedia. It's a generative pre-trained transformer. I can't find anything in the AI that is directing it to use actual, scholarly, or peer-reviewed, or even actual sources, versus "faking" them, as we've seen in the comments, above. AI takes no responsibility for this, nor do, it seems, its "handlers" (originators, developers). AI also cannot be "accused" of plagiarism or dishonesty, nor does it need to experience any consequences, unliked humans. It is responsible for nothing, and to no one. Humans, however, are.
Liz G 15 June 2023 AT 03:06 AM
I agree, but, the AI developers and those "riding the AI wave" (in the jargon of our IT people at our college and the Deans) don't really care to address such niceties. Harold Bloom was right, at least, about the internet, it was the beginning of the end.
B Shaw 13 July 2023 AT 05:07 PM
Chiming in to support this point! After doing some research, none of the articles cited in the screenshot appear to exist. This is a really important point to make to students. Based on the large amount of non-existent books and articles students (and faculty) have been requesting at my library, the fact that ChatGPT gives fictional sources is not widely known!
Rebecca Danielle Armentrout 01 May 2023 AT 10:05 AM
Does google translate when I am writing my own words in google translate to translate it in English need to be cited?
Dirk 09 May 2023 AT 09:05 PM
Hey, I was wondering, if students only used ChatGPT to rewrite their original work using more sophisticated academic terminology to express ideas more clearly and fix grammar and spelling mistakes. Do they need to cite this? If so, how do they need to do this?
Christa 22 August 2023 AT 09:08 AM
I came here for this guidance, too, and so far the most I can find is "acknowledge all functional uses of the tool (like editing your prose or translating words) in a note, your text, or another suitable location." So it seems that we need to guide our students to do this, based on our own preferences or a class discussion to arrive at a norm about it.
Christa 22 August 2023 AT 10:08 AM
I found this from below: Example:
“Rewrite the following text “Harry Potter and the Sorcerer’s Stone: Chapter 1” in 8th grade language” prompt. ChatGPT, 13 Feb. version, OpenAI, 12 Jun. 2023, chat.openai.com/chat.
To transfer this example into a possible way to cite editing/revision help from GPT that students are seeking, we might ask the student to document the editing prompts that they fed into the GPT to help them. It seems weird to put them in the Citations page though, because they'd be citing themselves, but that may be the thing to do?
Anon 15 May 2023 AT 11:05 AM
How would I cite ChatGPT or Google Bard if I were to use it for editing and brainstorming? What would I put in my bibliography? Every single prompt or just a statement of acknowledgement?
Sarah Pierce 12 June 2023 AT 04:06 AM
One thing we are seeing in class is that students are using ChatGPT to rewrite texts in easier-to-use language. We are discouraging them from summarizing/paraphrasing because that is not student-thinking. In this case, the prompt would include the source text (and could possibly be longer than the original essay in the works cited). therefore we have adapted the MLA citation to include the title of the source text rather than the source text itself.
E.g. Note: If your prompt includes a source text (e.g. the entire first chapter of Harry Potter), use the title of the source text in the citation instead of the text itself.
Example:
“Rewrite the following text “Harry Potter and the Sorcerer’s Stone: Chapter 1” in 8th grade language” prompt. ChatGPT, 13 Feb. version, OpenAI, 12 Jun. 2023, chat.openai.com/chat.
Patty Hude 07 September 2023 AT 10:09 AM
Hi! Why wouldn't the title of the source (the prompt) be capitalized in title case as is usual for MLA titles? (ex. “In 200 words, describe the symbolism of the green light in The Great Gatsby” is not in title case)
Kun Cao 10 September 2023 AT 07:09 PM
Hey MLA!
I was wondering if you guys have heard of AI Archives. It is a tool that saves ChatGPT, Bard, Claude conversations into a public URL. This is convenient for providing citations. Providing the prompt may not be sufficient (or just too long) given that the responses from Generative AI is non-deterministic. With a public URL, the reader is able to have access of the full context of the conversation used by the author.
Chicago Manual of Style recommends this tool!
Rachel Rigolino 16 September 2023 AT 07:09 AM
You say that a writer should cite ChaGPT if used as editing tool. Does this mean Grammarly should be cited and other tools of this nature?
Anuj Gupta 14 October 2023 AT 09:10 AM
Please provide examples of when a text is edited globaly for functional uses (clarity, summarizing, formality etc), what kind of acknowledgement statement to provide. Are end text citations and in text citations needed? If the ideas are an authors' own and AI tools like grammarly or Chatgpt are used to edit the tone globally, is an acknowledmenet statement good enough or does each edit need to be citext with in-text citations, which would make it too cumbersome in actual practice and risk making the authors' ideas seem like they were generated completely by the AI tool. Some guidance with examples on global uses of ChatGPT for different types of editing would be helpful. Thank you.
Loring Pfeiffer 20 October 2023 AT 08:10 AM
Yes, Anuj Gupta! I also would like to see an example of this type of acknowledgement!
Jeremy 25 November 2023 AT 11:11 AM
As would I like information about this,
Kenneth Sherwood 30 December 2023 AT 04:12 PM
Assuming this is still current MLA advice for citation, I'd encourage an edit. ChatGPT now includes a "share" link which generates a URL pointing back to the generated content in context with prompts. Thus I think your "Note 1" and the recommending of a Chrome plugin is no longer necessary.
Michele Hardy 12 January 2024 AT 10:01 AM
Hello, if students are permitted to use generative AI tools for prewriting purposes including generating topic ideas, brainstorming, and assistance with outlining, is a brief note of acknowledgement of this use within the assignment sufficient as attribution (similar to what is suggested if a tool is used for editing or proofreading)?
Or, would they need a Work Cited entry---i.e. "What are some controversial topics currently being debated in 2023?" prompt, ChatGPT, etc. etc.
Ed Caja 27 March 2024 AT 07:03 PM
Will this be coordinated with legal citation practice (Bluebook)?
Liam 29 March 2024 AT 02:03 PM
What kind of consideration is there for crediting models/Loras when it comes to custom generated images/responses? Frameworks like Ollama and Fooocus are making it easier and easier to mix and match FOSS generative AI components for personal use. How can we credit the authors of these components, and their source material?
Emily Connelly 05 April 2024 AT 02:04 AM
I am really happy to see such nuanced comments on this issue. I have been using generative AI in various test scenarios to see the potential use cases/pain points and as these comments speak to the complexities I think there will be in working with these technologies ethically. Our use of these should be acknowledged, our consent should be mandatory and separate from ToS and universities should require that disclosure in contracts and disclosure should be made whenever AI is deployed as a public facing point of contact.
Working with AI, I've fallen in love with Claude as a model and this conversation of citation is emerging more and more as a concern because, while he hallucinates, he also produces content for the projects we work on that I can easily use without adjustment. Our work is conversational to test the potential of the model so I am not going to accurately reflect the context of the response with the initial prompt since it is adjusted iteratively like a co-author's work. This is also true of creating visual works since the final prompt will be working off of a lineage of adjustments from an original prompt and these can be strings of texts too long to be practical in a source list.
Personally I do think AI disclosure at the beginning of the paper is important, both to acknowledge the tool and head off any accusations of hiding this and second because people need to be alerted to work that may need more consideration to accuracy due to hallucinations. Yes, they should do this anyway but we know how this goes and personally if it isn't a main citation I am not going to vet their entire reference list but if the citation comes from an AI I will.
Citation ideas that Claude and I are trying right now is to have a line under the author with a disclosure of use - we drafted this "With drafting assistance from Claude, an AI model developed by Anthropic. The author affirms that all content has been reviewed, edited, and supported with citations by the human author." and to then have a further citation in the references in the same formatting as a website source. In my work I've also archived a copy of the conversations referenced with a timestamp using a program like FireShot and have those available where I publish. In my case I want to share out the method and concerns but this would be no different in a way than if it were original data attached to an article.
The other reality is we are soon going to end up with AI stacks in our regular process. I just found, for example, a digital library AI that integrates with Zotero where I can have it fill in a table and answer queries about papers I store there which is accurate enough to be incredibly helpful in managing my thousands of thumb drives with academic articles named 284739w2.pdf within an hour and yes, I will be using that from now on. It is accurate to a degree but imagine if I didn't disclose that and how would I disclose that? The article is not theirs but they helped me sort them and summarize key points for reference during research.
The last point of this lengthy comment is that how we do this now builds how this will be. We need to make sure we use AI ethically in our work and part of that means being transparent with our use. If anyone reading this is an educator also remember how you work with AI shows students what they should do. If you don't talk about how you work with AI you lose great opportunities to normalize ethical practices. And I am so looking forward to how teachers find use for these tools and clever ways we can create engagement moving forward☆
Jennifer 08 April 2024 AT 06:04 PM
How do you deal with situations when you use multiple prompts to accomplish the task? Refining an original output is a common practice.
Bridget 16 July 2024 AT 03:07 PM
I'm also curious about this, especially because teaching students to refine prompts is an important part of effectively integrating genAI tools in a writing process.
Lisa 23 April 2024 AT 03:04 PM
What is MLA's guidance for students who use AI to translate their native language into English? Do they need to cite AI as the source of the translation? And does doing so violate course student learning outcomes that require students to "Use standard written English that includes a minimum of spelling errors, sentence fragments, comma splices, fused sentences, and agreement errors" or "Compose texts incorporating rhetorically effective and conventional use of language: clear expression, coherent ideas, purposeful word choice, complete development, and logical organization"?
Jordan 30 May 2024 AT 02:05 AM
This is very similar to my question below yours. This is a serious concern for teachers who evaluate writing.
Jordan 30 May 2024 AT 02:05 AM
I am curious about giving guidance to students using editing or translating websites that use AI, like "Wordvice" or "DeepL." The student will write a piece and then copy/paste it into the site, asking it to edit the tone, grammar, etc. This will often take a student writing and change it to the point that it is no longer the "student's" work. What sort of advice is recommended for this sort of situation?
Jacob Longshore 30 September 2024 AT 08:09 PM
One point that needs to be addressed: transparency. If a transcript of the conversation isn't included in a paper, readers will have no idea what the AI tool actually says. Quotes or paraphrases are worthless unless we have the original to compare them against; you could literally make up anything you wished to say, cite ChatGPT, and it would be impossible to verify or falsify. The most anyone could say is, "That doesn't sound quite right, but OK." That's not scholarship.
Depending on how AI is used, it could be a fruitful addition to research, e.g. when studying AI behavior. But the problem of accountability remains. In most any other case, we're given everything needed to track down the original source - which is impossible with AI conversations, since the same prompt can have different results. For the sake of consistency and transparency, include this in your guidelines.
Join the Conversation
We invite you to comment on this post and exchange ideas with other site visitors. Comments are moderated and subject to terms of service.
If you have a question for the MLA's editors, submit it to Ask the MLA!