Skip to main content
XCITE Center for Teaching and Learning

Department of University Writing AI Guidance

DUW Course-based Approaches to Generative Artificial Intelligence (GenAI)

Instructor Facing Content: This material is designed to assist you in considering A) the different possibilities for GenAI use in your classroom in conjunction with academic integrity statements; B) deciding on the "official" level of acceptance and use of GenAI for your classroom; C) learning of the UWP recommended implementations for GenAI in each course.

Unfettered Access to AI

The unfettered and unguided use of GenAI disrupts reading and writing because its product creates a situation where not all written words reflect the thinking and intention of a person. This makes offering accurate assessment difficult for instructors, especially if they attempt to ignore the reality that GenAI is available to their students. When left unaddressed, it becomes impossible for an instructor to accurately assess student work.

Academic Integrity Statements for Syllabi

It must be understood, as a baseline, that the unauthorized use of GenAI by students in their work falls under the existing definitions of “fabrication” and “unauthorized collaboration” that are used by SCAIP at UCR. This means that their unauthorized use constitutes a serious violation of academic integrity. It may not be “plagiarism” in the ways we are used to thinking of it, but its use is a violation of academic integrity when the outputs of GenAI are presented as a student’s own ideas.

Further, the use of GenAI is “unauthorized collaboration” when an instructor has not permitted the use of such software in their classes. It is, therefore, necessary for instructors to include robust Academic Integrity policies that directly address AI in their syllabi. Note, however, that asking students to “cite AI” is an inadequate response because AI outputs are non-reproducible and therefore obfuscate the purpose of academic citation. Below is one example of the relevant paragraphs of such policies:

“Academic misconduct can include facilitating cheating for yourself or other students, plagiarism, unauthorized collaboration, using apps or software programs to produce research results, and the fabrication of academic materials (including materials fabricated by Artificial Intelligence apps or programs).

Violations of academic integrity include, but are not limited to, using Artificial Intelligence apps or programs to generate text or ideas; using translators or translation software, including Artificial Intelligence apps, to translate text or ideas; copying text from any source (including search results, translation programs, chat bots, and the like) without proper attribution; using apps or other software to ‘revise’ your writing without permission, documentation, and attribution (including, but not limited to, Grammarly); turning in a paper written for one class in another; copying from a fellow student’s paper (including from their rough draft); using a paper belonging to another student from a previous quarter/year or another institution; improper paraphrasing; representing other people’s ideas as your own; and using or buying papers from online ‘paper mills’ and from in-person human writers. ” (Kathy Hardun, 2025)

AI Checkers Are Not Infallible

AI Checkers are unable to absolutely identify and locate AI-generated writing because the AI software leaves no identifiable trace data that can be quantitatively scanned as "AI generated." This is software that is not 100% and should not be relied on as the sole form of evidence of academic misconduct under any circumstances. At present, this is not a technology that the UWP can officially support for the following reasons:

1 See: Pindell, Nate. “The Challenge of AI Checkers.” University of Nebraska.  and, Myers, Andrew. 2023. "AI-Detectors Biased Against Non-Native English Writers." Stanford University: Human-Centered Artificial Intelligence. 15 May, 

  • False Negatives: there are too many false negatives generated and students can also be falsely accused of misconduct, and, of these falsely accused students, students, a disproportionate amount are English language learners and neurodivergent writers.1

  • AI Checker Accuracy: the exact accuracy of AI Checkers is skewed and inflated for marketing and investment purposes. Our student population at UCR, due to its high level of multilingualism, means that we might face an unusually large number of false positives and therefore should be especially cautious of the use of AI Checkers.

  • AI Humanizer Scrubbers: students can purchase an app that can "scrub" or "humanize" their AI output, which makes it more difficult for AI checkers to positively match work product to AI output.

  • UCR SCAIP: Our Student Conduct group does not currently recognize AI checkers as part of their case building.

  • Equity Concerns: Broader equity concerns relate to the frequency of false positives not only to multilingual students, but to neurodivergent students and speakers of multiple English dialects. Overall, we must understand that, much like how LLM outputs cannot be meaningfully cited, they cannot be meaningfully identified.

A Better Way to Deal with the Problem: An Explanation of the Closed / Conditional / Open Policy Language

One way to think of GenAI use in the college classroom is as “Closed,” “Conditional,” and “Open.”2 The use of this language allows the instructor of each class in each discipline to reflect carefully on its learning outcomes, assignments, and place in the broader curriculum to decide if or when the use of GenAI may be appropriate. The use of these terms also helps to standardize the place of GenAI in education for students and clarify what is and is not permitted. The UWP advocates for the use of these terms not only for standardization, but because the vocabulary here is not anti-technology and is open to change and adaptation in the future as tools change and the skills students come to us with change.

  • “Closed” means that the use of GenAI is not permitted, but instructors may choose to discuss GenAI and their reasons for “closing” the class by explaining the ethics and/or limitations of GenAI. Instructors may also assign reading about GenAI in these courses.
  • “Conditional” means that the use of GenAI tools is permitted for some assignments, tasks, or otherwise defined purposes in the class, but it is not permitted for everything. This may also include limits placed on which tools can be used and/or require careful documentation of how tools were used.
  • “Open” means that the use of GenAI tools is generally permitted in a class, and the class and instructor are open to all types of use and/or the use of all tools in their classroom space and/or the assignments for the class.

Below, please find suggested ways in which this vocabulary may be aligned with the course descriptions and learning outcomes for each course taught in the UWP.

DUW Ruling on Student GenAI Submissions

In no course within the UWP should material generated by GenAI be submitted for a grade, without acknowledgement that the material is GenAI-generated or as part of a larger assignment that requires students to cognitively address some aspect of the work's generation.

  • An acceptable example could be if an instructor asks students to utilize GenAI prompting to create a paper or assignment and then asks students to write an analysis and breakdown of this assignment.
  • An unacceptable example could be if a student uses GenAI to generate significant portions of their essay and passes this off as their own work.

In addition, citation of GenAI-generated text is highly problematic, since it implies that the use of GenAI is equivalent to research and the use of legitimate, human-authored sources, and should therefore be avoided.

2 These terms themselves, as well as the underlying policy formulation method, is heavily inspired by the work that has been done at Northwestern. See: “Use of Generative Artificial Intelligence in Courses.” Northwestern.

  • We are piloting the open, conditional, and closed GenAI methodology for Writ 10, 20, and 30. At present, Writ 001, 003, 005, 007 are all considered "Closed" for GenAI use unless consultation with Keith Vance (lashv@ucr.edu) takes place. Some experimental uses of GenAI may be warranted for ELWR courses, but those uses need to be vetted.

    In most cases, the learning outcomes of ELWR may not align with the use of GenAI for the following reasons:

    • reading comprehension cannot be bettered via reading summaries alone or artificially reducing the reading level of a document.
    • students must attain a reading comprehension level that is adequate to serve as a foundation for possible later Conditional use of GenAI tools.
    • an opinion must reflect one’s own thinking.
    • learning how to avoid grammar and syntax errors is not the same as having them corrected.
    • GenAI does not have personal experiences and, therefore, cannot assist with the development of authentic examples.

    Further, the ELWR classroom benefits immensely from a workshop-based environment. Peer-to-peer relationships and conversation must be encouraged, including multilingual conversation in the classroom.

  • Writing 10 is the place for introducing students to critical AI literacy, yet the classroom must be either Conditional or Closed because some of our learning outcomes can be readily negated by GenAI use. Using GenAI to generate text or ideas is generally inappropriate in the 10 classroom, and it is possible to fulfill the goals of Writing 10 in a Closed paradigm via careful assignment design.

    The learning outcomes of Writing 10 may be supported by the use of GenAI for the following reasons:

    • critical AI literacy supports the broader goal of assessing the credibility of information
    • Google NotebookLM, and potentially other AI summarizing tools, can support research goals and assist with citation management 
    • GenAI assisted search through the library databases can support research goals and is unavoidable in some databases
    • understanding GenAI in a critical way can support the goal of developing awareness of audience, purpose, and genre

    The learning outcomes of Writing 20 may not be supported by the use of GenAI for the following reasons:

    • development of individual voice and perspective cannot be assisted by GenAI
    • reading comprehension cannot be developed solely by reading summaries or items at an artificially reduced reading level
    • understanding one’s own perspective and purpose can only be done through an understanding of oneself and the world around us

  • It remains the case that the use of GenAI to generate text or ideas is inappropriate in the 20 classroom. It is also vital to consider the ways in which we cannot allow “making a coherent argument” to supplant concerns about development of voice and perspective; moving from Writing 10 to Writing 20, with GenAI, we must not lose sight of students’ sense of purpose, audience, and perspective on the subject. At the same time, the use of specifically research-oriented GenAI tools is appropriate in Writing 20 for the purposes of finding sources and managing citations, yet this cannot replace fundamental instruction in information literacy.

    The learning outcomes of Writing 20 may be supported by the use of GenAI for the following reasons:

    • GoogleNotebookLM, and potentially other GenAI summarizing tools, can be used to assist with research and citation management
    • GenAI assisted search through the library databases can support research goals and is unavoidable in some databases
    • library workshops, run by the librarians, now include robust content about the differences between AI “search results” and finding credible information
    • it is now impossible to teach research best practices without teaching critical AI literacy
    • GenAI can be used to create viewpoint diversity when there is none among students for the purposes of discussion

    The learning outcomes of Writing 10 may not be supported by the use of GenAI for the following reasons:

    • making an articulate and convincing argument requires an understanding of oneself and one’s perspective on the world
    • assessing the validity, credibility, and utility of information in regard to the argument one is trying to make cannot be assisted by GenAI

  • It is generally recommended that Writing 30 be Conditional, although Writing 30 has learning outcomes that can, if critical AI literacy is already present among the students or is taught early in the course, be broadly supported by GenAI use, and it therefore may allow for broader GenAI use than a Conditional WRIT 20. Therefore, if an instructor is confident in the critical AI literacy of a class, this is the place for Open use, provided GenAI is not used to supplant students’ own ideas or wholesale produce graded written work.

    The use of GenAI to produce ideas remains inappropriate in the Writing 30 classroom, while the use of GenAI to wholesale produce content remains inappropriate without careful assignment design. Open assignments need to be carefully scaffolded to safeguard students’ voices, original ideas, and critical thinking skills, while also requiring students to think through, document, and reflect on their use of GenAI tools. It is also appropriate for GenAI use in Writing 30 to be more heavily Conditional, as in 20.

    Learning Outcomes & Writing 30

    The learning outcomes of Writing 30 may be supported by the use of GenAI for the following reasons:

    • GenAI outputs can be the subject of analysis as a genre
    • GenAI can shift the genre of inputs in ways that may further 30 assignments
    • GenAI itself, in the ways it works and creates outputs, can be the subject of analysis
    • in its research capabilities, GenAI tools, such as GoogleNotebookLM and other GenAI summarizing tools, can be used to further the goals of Writing 30 in much the same way as Writing 20
    • GenAI assisted search through the library databases can support research goals and is unavoidable in some databases

    The learning outcomes of Writing 30 may not be supported by the use of GenAI for the following reasons:

    • GenAI does not possess, and cannot provide, critical scrutiny
    • understanding the nature and substance of meaning making is a human skill that must be developed within human beings
    • literary and cultural theory are human concerns

Work Referenced

Department of University Writing AI Guidance

Preparing your Document...

 
Let us help you with your search