USING CHATGPT and other AI LARGE LANGUAGE MODELS: Pace’s policy on the use of AI/LLMs such as ChatGPT is still evolving. As educators and students of writing we cannot ignore this emerging phenomenon—nor should we. ChatGPT, the most popular text-generating AI program, can be a helpful tool in teaching and learning writing; but it obviously has strong potential for abuse. Students are expected to submit original writing in their assignments. If you consult ChatGPT or another AI/LLM for any reason at all, you must state so, with specifics, at the top of the essay above the title. Example: “This paper was written with help from ChatGPT on drafting, editing, and proofreading.”
Papers submitted with an AI Usage statement will be accepted but evaluated with especially close rigor. As always, you should submit essays written in your own voice and style, expressing your own point of view. Essentially, this means that using AI/LLMs should be an early stage in your writing process—and not (except for proofreading) an endpoint.
Papers that employ AI/LLMs submitted without a usage statement will be considered plagiarized. Your instructor reserves the right to revise this experimental policy during the semester.
Feel free to share and/or adapt for use in your syllabi! If you use or modify this statement, please credit Dr. Vyshali Manivannan, Dept. of Writing and Cultural Studies, Pace University - Pleasantville (CC BY-NC-SA 4.0).
At first glance, OpenAI text generators, most famously ChatGPT, seem like infallible, labor- and time-saving writing technologies. However, even in their most advanced forms, they can’t think. Instead, AI text generators mimic and plagiarize the bodies of preexisting text that train them and predetermine their output, from the Internet to datasets pirating copyrighted books. They use statistics and probability to supply the next likely character in an ongoing sequence, “spelling” words, phrases, sentences, paragraphs, and essays.1
You can think of OpenAI like a powerful autocomplete function or an algorithmic version of the Infinite Monkeys Theorem: if an infinite number of monkeys bangs away at typewriters long enough, they’ll eventually reproduce the works of Shakespeare.2 In other words, AI can generate combinations of letters and punctuation marks that might seem original—but they aren’t. AI-generated texts derive from the language corpuses that trained them, many of which have been illegally obtained.3
AI text generators don’t “know” how to develop an original voice, style, thought, interpretation, or argument. They answer questions partially, inaccurately, and artificially. They can’t skillfully identify genre, resulting in subtle shifts between an article summary and a book sales pitch. They use incompatible organization schemas pulled from different genres and disciplines, incongruous disciplinary conventions, and inappropriate textual materials and statistics. They can’t reliably identify the sources that contributed to their output, creating issues around citation. They “hallucinate”: that is, they invent data and citations, and present them as true. And if you’re environmentally conscious, ChatGPT-3 uses an average of 500 mL of water for every 25-50 questions, as we face a growing water crisis.4
Writers still learning their craft might not notice when these issues occur without multiple rereadings and revisions. Current Writing Studies guidance is to evaluate texts on their own merits, regardless of what is known about the writer’s use of AI.5 This means that student writers using OpenAI must be able and willing to invest time, labor, and sophisticated writing and editing skill into recrafting an AI-generated text into a submission-ready assignment that demonstrates awareness of genre, disciplinary conventions, style, and citation practices. This requires an extremely careful and critical editorial eye. Failure to “humanize” or catch incompatible writing conventions or styles in AI-generated texts is likely to flag them as completely written by AI.
AI-generated essays actually magnify common problems in student writing: uncertainty about genre conventions and appropriate style; confusing organization; inconsistent paragraphing; generalizations; logical fallacies; repetitiveness; flawed research and citation; inconsistency with your voice in class discussion and on other assignments; and problems addressing the assignment instructions.
Writing-enhanced courses aim to teach you that writing is an activity of thought, a skill that goes beyond correct grammar and mechanics.6 AI text generation doesn’t encourage the learning of reading and writing, let alone the cultivation of the thoughtful inner readerly/writerly voice that you’ll eventually need, whatever your career path.
In short, ChatGPT is not a time-saving tool unless:
❍ You already possess mastery of the target genre, writing style, research methodology, and material
❍ You know how to write a high-quality prompt for ChatGPT that precisely and specifically summarizes your assignment directions, target genre, writing style, and research criteria
❍ You know how and where to integrate your own critical thinking, interpretation, and analysis based on your class discussions and previous assignments
❍ You are a skilled source- and fact-checker and line editor and will verify that the information is accurate and locate the original sources on your own to provide attribution
❍ You are expert in revision
If you possess the above skillset, AI text generators can be helpful in the prewriting and paraphrasing process. However, like any tool, you have to use it with competence and care. The 4C/MLA Task Force on AI and Writing offers initial guidance for using and evaluating AI-integrated scholarly and creative writing. It’s the responsibility of the writer to center their own original thinking and contributions when using AI for brainstorming or organization; ensure accuracy and substantive quality (not merely grammatical correctness) and remove offensive or problematic statements in their work; do extra research to attribute ideas to scholars uncited in the AI output; and disclose the specific ways AI was used and cite the system, dates, and prompts in their documentation; triple-check for accuracy.7
Restorative justice pedagogies acknowledge that external pressures (work, health, personal crisis) can reduce your available writing time; self-disappointment (“I’m bad at it”) can make you writing-avoidant or ashamed to meet with me; GPA concerns (financial aid, honors programs, specific majors, athletics) might make you susceptible to problematic “get good grades quick” schemes. If this sounds like you, swing by for a coffee chat. I’m open to hearing your concerns, whether material or emotional, and working to ensure an equitable learning environment for everyone.
1MLA-CCCC Joint Task Force on Writing and AI. Working Paper: Overview of the Issues, Statement of Principles, and Recommendations. hcommons.org.
2Morrison, Aimée, keynote. Everything I Need to Know about ChatGPT, I Learned from My Students. 2023 Festival of Teaching and Learning. University of Alberta/Zoom.
3Estes, A. (2024 Jan 18). A New York Times Copyright Lawsuit Could Kill OpenAI. Vox.
4DeGeurin, M. (2023 May 10). Training ChatGPT AI Required 185,000 Gallons of Water. Gizmodo.
5MLA-CCCC Joint Task Force on Writing and AI. (2024). Initial Guidance for Evaluating the Use of AI in Scholarship and Creativity. hcommons.org.
6Association for Writing Across the Curriculum. (2023 Jan). Statement on Artificial Intelligence Writing Tools in Writing Across the Curriculum Settings. wacassociation.org.
7MLA-CCCC Joint Task Force on Writing and AI. (2024).
The use of generative artificial intelligence tools such as ChatGPT is permitted for limited and specific tasks, and may even be required for some assignments. Assignments using AI would compare and analyze the student's own work with AI-generated material. Aspects of your coursework that relate directly to the learning objectives of the class should not be carried out with AI.
The goals for this course include ways of developing your own tools for noticing cultural and gestural visuals. This is because I would like to monitor your progress in these areas and help you improve in them.
AI can be used to brainstorm ideas about films and text we are studying, and you may experiment with AI tools for outlines and in other areas, but the final project text must be entirely your own work. AI may not be used in quizzes or tests.
Any use of AI to assist you in your coursework must be fully transparent. This will require a citation or a disclosure statement. For information on how to do this, consult the citation section of Pace's Student Guide to Generative AI.
The following cautionary note appears in the guidelines for the use of generative AI tools at Harvard University: "AI generated content can be inaccurate, misleading or entirely fabricated (sometimes called hallucinations) or may contain copyrighted material. You are responsible for any content that includes AI-generated material." Given this truth about generative AI, always read over and fact-check its output.
I encourage students to use AI tools (Copilot, ChatGPT, Grammarly, etc.) to support their writing ethically and constructively. These tools can be helpful for:
Brainstorming ideas
Exploring directions in writing when you are stuck
Getting help with citations and formatting
Correcting spelling and grammar
I use AI tools myself, and I believe learning how to use them wisely is part of preparing for the future. However, you must acknowledge any use of AI in your assignments to maintain honesty and ethics in academic work. You may not submit AI-generated work and claim it as your own. If I suspect that you are doing so, I will ask to meet with you one on one so you can explain your work and each source to me. If you cannot to my satisfaction, I will be forced to fail the paper.
AI is a wonderful gift. But if you let it do your work for you, you cheat yourself out of developing skills that will give you a competitive advantage in your career. People who lack strong writing and editing skills—who cannot evaluate tone, audience, or appropriateness—will not benefit as much from AI writing tools. In an AI-driven world, human critical thinking and originality in writing will continue to become even more important.
If you have used AI in any way, please include an AI Assistance Statement at the end of your paper, after your conclusion or references. This statement should briefly describe how you used AI tools. Here are three examples of acceptable wording:
AI tools were used to help brainstorm ideas and organize the structure of this paper.
I used AI to suggest improvements to sentence clarity and grammar.
AI assisted in formatting citations and checking for consistency in APA style.
If you have questions about what counts as appropriate use or how to phrase your statement, feel free to ask me in person or by email.
Do you have a syllabus statement that you'd like to share? Please email kpradt@pace.edu to have your contribution added to the guide.
President Krislov's take: Why ChatGPT Makes Me Hopeful—Not Worried—For the Future of College and Careers (March 13, 2023)
Three Pace professors chime in via the Pace website: ChatGPT and the Future of Education (April 4, 2023)
Article on the work of computer science professor Christelle Scharff: Leveraging African Fashion for Addressing AI Bias (July 13, 2023)
Report on a talk given by chemistry professor Elmer-Rico Mojica: Educators Urged To Set Parameters When Using AI Tools For Teaching (August 14, 2023)
Philosophy professor James Brussard weighed in on the ethics of AI in academia: An Ethics Expert's Perspective on AI and Higher Ed
Also take a look at our companion guide for students, Student Guide to Generative AI, which contains more and different AI-related resources.