Citation Styles and Management Tools Guide

NOTE: Information about using and referencing ChatGPT and other generative AI tools is still being developed and will continue to be updated.

When an assignment allows AI tools, you must acknowledge and/or cite any AI-generated material that informed you work. This includes if you use an AI tool to help write or structure your paper, even if you do not quote or paraphrase its content. You must acknowledge your use of the AI tool to provide transparency to your reader. Using AI to generate content without proper attribution can qualify as academic dishonesty.

ACE your usage of AI by:

• Acknowledging your use of an AI tool in your research or assignment
• Citing the tool in your Works Cited
• Evaluating the information given to you by the AI tool, verifying its sources and checking for errors

ACE acronym developed by STEM Librarian Renée Walsh

This is the best approach when you've used AI in support of the research process, but haven't included any specific output from the tool that you could cite. Acknowledgements in research papers are a place to recognize a variety of contributors to your work. This can include academic mentors/advisors/supervisors, funding bodies, technical supporters, colleagues/collaborators who contributed intellectually (but are not co-authors), and even personal supporters. It can also be used to identify when and how AI tools were used in the research process.

Remember: before using an acknowledgement, be sure to confirm with your instructor if the use of Generative AI tools is allowed for your assignment. If permitted, your instructor may also have a specific way they would like you to acknowledge the use of AI tools in your work.

Further help building an acknowledgement statement can be found on the University of the District of Columbia's Acknowledging and Citing Generative AI: Acknowledging AI Use page.  

While an acknowledgement broadly recognizes contributors to the research, attribution is the process of giving specific credit to the source of information or ideas used in one's work. The specificity of the source of information is a big part of the issue when trying to cite AI generated content.

Large Language Models (the technology that powers generative AI tools) recognize patterns and relationships in language. Through pattern recognition the system can generate predications about what word might come next. Notice we'd said patterns and prediction, but we haven't said anything about thinking or searching. These tools are not searching for sources of information and they're not thinking of new ideas, they're stringing together a plausible set of sentences gathered and remixed from the text it was trained on.

Since the technology is generating plausible strings of words each time, it can be almost impossible for someone to get the same output twice. Citing the source of the information the AI used to generate its output is much more stable. Perplexity is AI tool that links to some of the sources it pulls information from, making it slightly better suited for a research role, but those sources need to be carefully checked as they may not contain the information Perplexity claims it found there.

Given the nature and limitations of the technology, if possible you should cite the original information sources the AI tool pulled its information from, instead of the tool itself. However, if you've verified the accuracy of the information, but aren't able to locate the exact original source of the information, you should cite the AI tool. You should also cite AI if you quote summaries or synthesis that the tool has uniquely generated.

Generative AI tools powered by large language models are rapidly proliferating and developing. Any guidance about using these tools to support citation formatting, style transformation, or in-text citation review are current to the last date of editing on this page.

Can AI-powered chatbots check the formatting of my citations?

Yes, however, there are two important limitations:

1. A chatbot will only offer feedback on the specific formatting guideline it is prompted for. Generic prompts to check your works cited may miss crucial errors. More specific prompts, such as checking your works cited for proper use of italics according to a particular style type, will produce higher quality results.
2. A chatbot will only offer feedback according to the most current edition of a style guide and the most popular style types. If your assignment requires a prior edition (i.e., MLA 8 instead of MLA 9) or a less common citation style (i.e., ASA), you will need to provide the chatbot with reinforcing documentation to guide its feedback. Attaching a style guide text along with the works cited document you wish to check will produce a higher quality output in these cases.

Can AI-powered chatbots transform my citations from one style type to another?

Yes; however, the same two limitations apply in this use case as above. A quality output in the new citation style requires a careful eye for mistakes and very specific prompting. If you are translating into a less common citation style, you will need to provide the chatbot with reinforcing documentation to guide its feedback.

Can AI-powered chatbots check the formatting of my in-text citations?

Not reliably. To date, chatbots will provide general guidance on in-text citation, but when asked for specific revision assistance, they will garble the page reference or in some cases fabricate errors. This work is still best done by dedicated writing assistance tools, though those may not be free to use and still need human verification.