Statement on Generative AI and Writing
September 2023
It is the University Writing Program’s position that genAI is a powerful and productive development in the long history of literacy technologies, and that the teaching of writing should include methods of integration and collaboration with these technologies while also interrogating the rhetorical and ethical dimensions of genAI.
In conjunction with this position, the University Writing Program also acknowledges the following:
- GenAI is not a replacement for the act of writing and its positive effects on learning. GenAI is an additional dimension to and not a substitute for writing.
- GenAI is not a replacement for understanding what effective writing is and how writing functions.
- GenAI applications are designed by humans, and as such, carry biases and a preference for the mean that can hide diverse voices.
- The technological, social, financial, and legal dimensions of genAI will continue to change over time, so potential use cases will also change. Reliance on any one tool is shortsighted.
- Much like all literacy technologies, genAI is neither an inevitable nor ubiquitous development, and contexts and expectations will change. Thus, writers should be versatile and adaptive, and most importantly, actively contribute to literacy technology development.
- Current genAI is prone to hallucinations, or fabricating evidence, even when prompted with a specific source.
Responsible and Ethical Use of GenAI
Artificial Intelligence (AI) has been an important development in computer technologies, and advances in generative AI (genAI)—AI that can create something new—has been used in the past decade in business and journalism. Developments in genAI, especially those applications that use Large Language Models (LLMs) using billions of parameters, have come in leaps and bounds due to the free sharing of studies from both academia and industry. Nevertheless, it still took some by surprise when, in November 2022, OpenAI publicly released for free ChatGPT, a text-to-prose transformer genAI, that produced extraordinarily articulate writing in a variety of genres, from essays, to computer code, to poetry, to business plans.
The public availability of these powerful applications has provided opportunities for many to explore, play, and collaborate in new ways with literacy technologies. For those with access, the potential to address skill inequality has shown great promise. Of course, this has also led to profound concerns in higher education about genAI’s negative effects, including student use without guidance.
And yet, genAI has already been integrated into Google Docs and Microsoft Word, two of the most popular word processors writers use. The integration of these technologies will continue to proliferate, so that many common applications will eventually have genAI assistance, much in the same way that literacy technologies have developed to include design and formatting elements with little input from a user.
The University Writing Program supports responsible and ethical use of genAI assistive writing technologies by writers and teachers.
For writers, responsible and ethical use means the following:
- Before using any such technology, review the policies and constraints of the rhetorical situation. As a student, this means abiding by instructor policies for assignments communicated in the syllabus, assignment, or verbally. For publications, that means reviewing editorial policies. For business, that means contacting your managers and reviewing company policies.
- It is valuable and ethical to acknowledge in writing what assistive technologies you used and for what purpose. Acknowledgement statements have always been important in academic and professional writing, so it’s a good habit to use them in other genres and contexts as well. Some examples might include the following:
“GPT-4 was used for help with wording, formatting, and styling throughout this work.”
“ChatGPT was used to help organize early drafts of this project.”
“Portions of this articles were revised based on suggestions made by Grammarly.”
“Google Bard was used to generate the idea about X in early drafts.”
- Review the End-User-License-Agreements (EULAs), privacy policies, and other use documentation so that you understand your rights, the rights of the developer, and the use of data you provide to a company. Never input private, propriety, or protected data to a third-party website without understanding these policies.
- You are responsible for writing that you submit. That means you are accountable for genAI prejudices, misinformation, or fabrications in the texts you submit. A genAI cannot be an author/co-author in the traditional sense of the word, and it should not be cited as such (much as you wouldn’t cite Microsoft Word or Google checking these policies.
- Take advantage of opportunities to improve your writing through practice. Such practice might involve interacting with genAI to help you get immediate feedback, but genAI should not be used as a substitute for the process of drafting, revising, and reflecting.
- Make suggestions and offer feedback to application developers to address biases or other potentially harmful results while using such technologies. Do not be a passive user. Docs.)
For Faculty assigning writing tasks, responsible and ethical integration means the following:
- Acknowledge that we are aware that genAI exists and outline clear policies about when and how it can or cannot be used in an assignment or task. Blanket syllabus statements forbidding or giving carte blanche are not as helpful as outlining the ways genAI might be used for specific tasks or giving specific directions not to use genAI for certain parts of the writing process. This should include instruction for writers about how to acknowledge the assistance they received on a text.
- Create time to discuss the ethical implications of using genAI for assigned writing tasks. That means reiterating issues related to inputting protected data for research tasks, responsibilities of ownership for what the writer submits, and importance of context and policy inasmuch that different audiences and purposes might expect or forbid the use of such technologies.
- Encourage students to take an active role in their technology use. Build in discussion or assignments related to a course about genAI’s potential impact on your specific discipline or a course topic. While we recognize that it might not be possible to address every potential impact, this statement is a public document and can be a good resource for students in order to get them to begin thinking of responsible and ethical uses as well as larger environmental, financial, and bias concerns of genAI.
- Do not rely on applications that purport to detect genAI use (e.g., Turnitin, and especially, genAI itself) as they are imperfect and produce just as many false positives as false negatives.
- Read writing generously and provide feedback that rewards diverse voices. Recognize that assignments that mirror common writing tasks often lead to the type of writing that the LLMs were trained on. Therefore, encourage new approaches. Multilingual writers, in particular, might turn to genAI if feedback is overly critical of errors.
- Discuss why you are assigning a particular writing task. If students know why you want them to write (e.g., writing-to-learn, writing for an audience, writing to approximate genre features), they are more likely to attempt the task with this intention.
The University Writing Program and University Writing Center are happy to consult with individual faculty or departments on strategies for integrating genAI into writing assignments.
We would also like to share the following additional resources:
- Artificial Intelligence Usage at the University of Denver: Recommendations and Resources for Faculty, Staff, and Students
- MLA-CCCC Joint Task Force on Writing and AI Working Paper: Overview of the Issues, Statement of Principles, and Recommendations
- Statement on Artificial Intelligence Writing Tools in Writing Across the Curriculum Settings
- TextGenEd: An Introduction to Teaching with Text Generation Technologies by Tim Laquintano, Carly Schnitzler, and Annette Vee