Student at desk black and white
Collection

How Can I Promote Academic Integrity With AI?

Generative artificial intelligence (AI) has sparked renewed concern about academic integrity.  Here, we recommend resources to help you consider ways to promote academic integrity in your courses, whether you choose to encourage or discourage the use of AI.

Updated December 2024
Jess Taggart headshot
Assistant Director & Assistant Professor
Office of the Executive Vice President and Provost
View Bio

Promoting Academic Integrity in Your Course

Cornell University

Cornell University describes three concrete strategies for promoting academic integrity in your courses, with special consideration then given to AI.

Headshot of Jess Taggart
Jess Taggart
This is an excellent place to start as you consider how to foster a learning environment grounded in trust and academic integrity. The suggestions provided are immediately useful and relevant across disciplines.
View excerpt
Seemingly every year, there is a new platform or website that prompts us to think about how we are assessing student learning and how we can ensure that students are demonstrating what they know with integrity. What follows are several strategies for promoting academic integrity, as well as some specific information on Artificial Intelligence tools (or AI tools, e.g., ChatGPT) that may help inform how you assess learning and create a learning environment that fosters trust and academic integrity.
Was this resource helpful?

What Counts as Academic Fraud at UVA

UVA Honor Committee

What qualifies as academic fraud at UVA? The Honor Committee website provides examples and contact information to learn more.

Headshot of Jess Taggart
Jess Taggart
Each institution may approach academic integrity slightly differently, particularly when it comes to AI. Knowing what UVA’s expectations are is essential.
Was this resource helpful?

Citing Generative AI

UVA Library

This University of Virginia Library Guide provides guidance on how to cite generative AI models in three popular styles: APA, MLA, and Chicago.

Headshot of Jess Taggart
Jess Taggart
If you plan to allow students to use generative AI and want them to cite their work, this is a useful website to share.
View excerpt

Citations are a critical component of academic writing. They allow you to give credit to the original source, whether you are using a direct quotation or paraphrasing someone else's research. Citations give the reader directions to specific passages in a text, draw attention to prominent authors in the field, and can give a historical overview of research around a particular topic.

Different disciplines have different norms for how information is cited and they prioritize different parts of the original source. Citation styles give the reader a way to quickly access the information you are citing and can give a quick overview of the body of work you referencing. Some styles prioritize the year so that the reader can see how the scholarly conversation you are referencing has evolved over time while others include the page number so that a reader can refer to the exact phrase you are discussing. It is important to remember that different styles use a different combination of punctuation and typography. 

After ascertaining whether the use of AI generated content is permissible, determine how to reference the material in your work. Make sure to include citations both in-text and in the references at the end of your work. Since AI tools are relatively new, it's possible that your required style hasn't updated its rules around AI citations. The purpose of citations is to provide a guidepost to the source material so, in the absence of formal guidance, be as transparent as possible about how you obtained the information.

Was this resource helpful?

Documenting Your Use of Chat GPT: A Guide for Students

Kimberly D. Acquaviva

This 5-page guide composed by Kimberly D. Acquaviva walks students through how to document their use of ChatGPT on assignments—and shows them why fact-checking the content curated by ChatGPT is a must.

Headshot of Jess Taggart
Jess Taggart
This guide is an excellent example of how to provide students with clear information about how they should cite the use of generative AI in assignments. As an instructor, consider how and when you would like students to document and disclose their use of AI in assignments and try generating your own guide.
View excerpt
Artificial intelligence (AI) tools like ChatGPT can help you brainstorm ideas for essays, policy briefs, draft outlines, generate rough drafts, analyze texts, and compile annotated bibliographies. While tools like ChatGPT are helpful, they are only tools. Even the world’s best nails are useless without a hammer -- and without a person who knows how to swing it. Library research skills are the hammer you’ll need in order to use ChatGPT effectively. Whenever you use AI tools to write a paper, essay, article, or policy brief, you should include an explanation of how you used the tool, the prompts you used to generate output, and how you fact-checked the output.
Was this resource helpful?

Our Obsession with Cheating is Ruining Our Relationship with Students

Marc Watkins

In his blog post, Marc Watkins shows, with screenshots, how AI-detection tools may not detect cheating, and he pushes instructors to consider how our drive to prevent cheating should shift to a focus on creating a culture of trust.

Headshot of Jess Taggart
Jess Taggart
I appreciate how this blog post provides a valuable opportunity for instructors to reflect on what they hope to accomplish with AI detection software, and what the cost of those efforts may be.
View excerpt
As the technology to generate writing with AI ramps up, so does the production of AI-detection tools that attempt to curtail “cheating” by using them. But the bevy of AI detection tools on the market aren’t keeping up with the speed of AI’s development. Because they are using older versions of GPT as the main component of their detection mechanism, they often misidentify human-composed text as AI-generated or confuse AI-generated text as human-generated. And while OpenAI is working on a watermarking scheme that uses cryptography to mark text generated by a transformer, even their lead engineer admits this would be easy to defeat by simply copying a generated output into a different language model.
Was this resource helpful?

Want to recommend a resource to add to this collection? Send us an email.