Although it has been only two years since the wide release of ChatGPT, Generative Artificial Intelligence (GenAI) I technologies are becoming ubiquitous. Familiar technology companies like Google, Microsoft, Meta, and Amazon already use large language models to design new writing platforms and technologies. In most cases, people will use AI to augment their writing processes rather than as a replacement for human writers. The challenge for users will be to make decisions about what AI can do effectively (basic summary and general description) and what may still depend on human agents (attribution, evaluation, and judgment).

Policy Options for Incorporating GenAI in Undergraduate Courses

The Provost’s office established initial recommendations for GenAI policy focusing on ChatGPT, the most ubiquitous platform at the time. The three recommended syllabus statements allow individual instructors to embrace, allow, or prohibit the use of ChatGPT and other GenAI tools.

For instructors who wish to embrace ChatGPT

Artificial intelligence (AI) language models, such as ChatGPT, may be used for any assignment with appropriate citation. Examples of citing AI language models are available [or provide an alternative reference appropriate for your class]. Students are responsible for fact-checking statements composed by AI language models.

For instructors who wish to allow limited usage of ChatGPT

Artificial intelligence (AI) language models, such as ChatGPT, may be used for [assignment types A, B & C] with appropriate citation, but not for [assignment types D, E & F]. Please discuss your situation with me if you doubt whether you use AI language models appropriately in this course. Examples of citing AI language models are available [or provide an alternative reference appropriate for your class]. Students are responsible for fact-checking statements composed by AI language models.

Documentation guidelines for ChatGPT

Students must document their use of GenAI in courses where it is allowed. Different stylesheets have different expectations for how AI-generated language should be documented; the difference between APA style and MLA style is especially interesting.

American Psychological Association Style

APA treats text generated by AI tools as personal communication, much like an email to an author. The in-text citation below follows the author-date convention typical of APA.

ChatGPT offers the following description of the APA style: “The APA style refers to the formatting and citation guidelines established by the American Psychological Association (APA). It is commonly used in academic writing, particularly in the social sciences such as psychology, sociology, education, and other related fields” (OpenAI, 2023).

The entry in the References section of the document would list Open AI, ChatGPT’s owners, as an author.

OpenAI. (2023). ChatGPT (July 20 version). http://chat.openai.com

Modern Language Association Style

MLA treats the user-generated prompt as the title of a document and does not credit ChatGPT (or other AI) as an author. The in-text citation and reference below follow the preferred conventions of MLA, which treats ChatGPT as a source without an author rather than an agent that generates text.

“MLA style, short for Modern Language Association style, is a commonly used citation and formatting style in academic writing, particularly in the fields of literature, arts, humanities, and some social sciences” (“What is MLA Style?”).

On the Works Cited page, the entry would have no author listed.

“What is MLA Style” prompt. ChatGPT. 20 July version. Open AI. 20 July, 2023. chat.openai.com/chat.

In each example above, the AI-generated text is offered as a direct quotation set off by quotation marks. In addition, the text appears highlighted when pasted from the ChatGPT interface. This highlighting is an initial attempt to help readers determine that generated text comes from a large language model.

While rules for direct quotation are established, indirect quotations, paraphrases, and summaries can introduce gray areas for students. The best advice for students is to always document the use of GenAI, even in cases where a parenthetical reference may not be required. Any failure to document Generative AI use violates the Code of Student Conduct.

Recommendations for Instructors

Enter your writing assignments into Microsoft CoPilot and other models to see what is generated

The University of Minnesota has acquired limited commercial licensing for MicroSoft CoPilot with enterprise protection. The licenses afford faculty, instructors, and students access to GenAI technology without concern that their inputs could be used for future training. The text generated will be based on statistical probabilities of matching responses and will follow the language parameters upon which it is trained.

When submitting a prompt to a GenAI, consider the elements where generated text successfully meets assignment expectations and where it is unsuccessful. For example, users may notice that the form of the in-text citation is correct (author, date) but that the content doesn’t refer to an actual source or author. Similarly, a GenAI-generated reference page may be organized alphabetically but may not follow APA rules correctly (or consistently) for capitalization and typography. Assignments can be modified to emphasize the skills and activities that require the most student labor and deemphasize the summary skills where AI use might be tempting to students.

Help students to understand the limits of GenAI and its tendency to errors

It is vital to help students recognize the limits of generative AI. While digital detectors have difficulty distinguishing between AI- and human-written prose, experienced human readers often identify the errors and hallucinations produced by generative AI. Novice readers are less likely to catch these errors and may, unfortunately, be more likely to be persuaded by the authoritative style produced by GPTs.

Teach students strategies for prompt design and effective narrowing

Initial outputs from ChatGPT and other tools are often exceedingly conventional and lack relevant details. However, by using a process of prompt refinement, users can generate text that comes closer to serving the purposes and audiences intended with its output. ChatGPT offers six basic strategies for improving results from initial queries. In addition, many technology companies are developing Application Program Interfaces (APIs) to assist ChatGPT with complex tasks like statistical analysis and complex mathematics. Such ‘bolt-on’ technologies can improve the accuracy of responses and limit the generation of “hallucinated” results.

Remind students of alternative or better resources.

Students sometimes use GenAI as a virtual tutor or research assistant to generate ideas or learn strategies for beginning an unfamiliar assignment. Remind students that their instructors and TAs are subject matter experts who may have tested and proven methods for explaining materials and concepts in many ways. The information, analogies, summaries, and comparisons generated by actual experts rarely contain the kinds of errors and omissions common to LLMs like ChatGPT.

Similarly, while ChatGPT can generate keywords to guide literature research, librarians and subject matter experts can help students use library tools and interfaces, controlled vocabulary searches, Boolean operations, and limiters to generate more effective and accurate results. And if students are interested in conversation or feedback about their work in progress, they will find consultants from Student Writing Support to be much more attentive to their unique needs and strengths.

Ultimately, while GenAI tools often appear to present authoritative statements and practical advice, they are no substitute for human professionals' assistance.