Citing Generative AI

Why cite generative AI like ChatGPT and other results from large language models (LLMs)? 

Remember that according to Brandeis's Academic Integrity guidelines, "Student submitted work must be the product of that student’s own thought or study with proper attribution and citation," so if you are using an AI tool, whether ChatGPT, DALL-E, or any other generative tool, you will need to cite it appropriately, as the information needs attribution.

Most plagiarism is unintentional, but it's still considered an infringement of academic honesty. See Section 4 of Student Rights and Responsibilities for more information. 

You can avoid unintentional plagiarism by citing your sources correctly.

But why? Remember that along with proving that you're not plagiarizing, citation also provides two other major purposes in academic work: giving credit to the author or creator, and making sure that the next scholar reading your work can locate the sources used in your research. If you use a generative AI tool, providing a citation helps to give credit to where an answer came from--both as evidence of what kind of LLM is the creator, and helping others understand what source you're consulting in your research.

Jump straight to our style-by-style outlines for Chicago Manual of Style, MLA, APA, and other major styles (coming soon, as they're updated) using the left-hand navigation tabs.

Currently Zotero and EndNote do not have an option for generating a citation for AI-generated content, so you will have to hand-write your citations. 

Before you cite your generative AI source, make sure that you double check the following list: 

• Make sure you're allowed to use generative AI or an LLM in your assignment. If you're unsure, consult your professor.
• Fact-check any content or sources that the generative AI or LLM provides for you.
  • This includes finding and reading any provided citations, as well as fact-checking any provided information, which may be lacking context, be taken from an accurate piece of research but cited incorrectly, or wholly made up. Generative AI and LLMs are still consistently providing false or contextless information. This is sometimes referred to as "AI hallucination."
• Provide context in your research indicating that you used generative AI or an LLM to gather information, including text prompts you used and how you used the tool.
  • This may not be in the body of your paper, but in a footnote or appendix, depending on the style. Check your specific style using the tabs on the left-hand side of this page. Consult your professor if you're unsure how to include this information. 

We've included guidelines from the major citation styles below but be aware that this world is changing quickly, and citation styles may lag behind. If you're not sure how to cite something, reach out to your professor or instructor. 

Intellectual Property & Copyright Considerations:

Current intellectual property and copyright laws are still catching up to AI and LLM usage. The US Federal Government is working on creating legislation about copyright and AI, and some pieces have gone into effect, but others are still in flux. Please consider the source or usage of your LLM/generative AI output and input carefully.

Privacy:

Many AI tools will also often take and use your prompt data to help improve the software.

Assume that any prompts and information you put into the tool will become part of the tool, and therefore public.