1. The Honor Code ›
  2. AI Recommendations for Instructors

AI Recommendations for Instructors

For more: check out ND Learning's Teaching in the Age of AI resource website

Working Group Recommendations (May 2023)

  1. Remind students that ChatGPT and related tools do not meet the standards normally expected of academic or professional work. The lack of citations, the inability to reproduce responses to the same prompt, the difficulty of distinguishing fact from disinformation, the lack of accountability for the ethical and equitable sourcing of information, and outright “hallucinations” are among the known issues. Learning to think and write critically and to use genuine sources of truth and knowledge form the hard but essential core of a liberal arts education, and are the most important skills required for success in a rapidly changing world.

  2. Be explicit with students about your expectations regarding the use of ChatGPT and related AI technologies for assignments, exams, and in the classroom, and be clear that engaging in unauthorized use of generative AI will be considered a violation of the Honor Code.

  3. Consider adding a specific statement in your syllabus and on Canvas regarding the use of AI technologies.

The following sample statements may be helpful in crafting statements tailored to the learning goals of specific courses:

    • Before beginning your studies at Notre Dame, you signed the Honor Code pledge, affirming that you would not tolerate or participate in academic dishonesty. My fair assessment of your learning in this course requires me to review your work as completed in accordance with our Honor Code. To be clear, for the purposes of this class, you may not engage in unauthorized collaboration to complete any work for the course, and you may not use ChatGPT or other AI composition software unless it is permitted for the assignment.
    • AI composition software (like ChatGPT) may not be used for assignments in this course unless specifically stated in the assignment. Using these tools without my permission puts your academic integrity at risk. When this use has been permitted, the content generated or significantly modified by technology means must be properly attributed.
    • The use of AI technologies (like ChatGPT) is generally acceptable in this course, but, like any other source, must always be attributed in a manner that can be reproduced by the reader. An example of “reproducible” attribution includes the date accessed, the web address or URL accessed, and a description of the prompt. When using generative AI technologies as a source of information, you are responsible for assessing the quality, completeness, and accuracy of the cited information.
    • The use of AI technologies (like ChatGPT) is generally acceptable in this course, but, like any other source, must always be attributed in a manner that can be reproduced by the reader. Moreover, such writing aids may be used in different ways at different stages of the creative process. Therefore if you use any of these systems to support your work where assignments specifically allow it, you will be required to describe the uses and process. Remember, when using generative AI technologies as a source of information, you are responsible for assessing the quality, completeness, and accuracy of the cited information.

  1. Become familiar with ChatGPT and related tools.

    • Engage with colleagues, students, and others in dialogue regarding the use and misuse of generative AI in your academic disciplines and the professions in which your students engage. These technologies are changing rapidly with large commercial investments. They will likely become pervasive, ubiquitous, and embedded within the common applications used by students, scholars, and professionals of many types. What this means for the high standards we hold for teaching, scholarship, and academic integrity is a vital conversation to which we invite your thoughts and advice.
    • Use ChatGPT yourself and put your assignments into the system to see how it responds. Become familiar with the tools and how your disciplinary community is understanding them. Impacts of these technologies on the disciplines vary as to how fast and seriously they will impact specific professions, academic disciplines, and pedagogical approaches.
    • Be aware of the limitations of currently available AI detectors. The detectors found online are not currently reliable enough to use as the basis for an accusation of academic dishonesty. They may, however, serve to draw an instructor’s attention to a piece of writing and nudge them to take a closer look, but there are too many false negatives and false positives for us to depend on them alone. A better strategy is to put your own prompt into ChatGPT, even 2-3 times, and compare the output to the student’s work. Or, check the factual claims and the bibliography / references in the text, since ChatGPT will often fabricate these.
    • Also, remember that ChatGPT does not know whether it wrote a particular piece of text; if you ask it if it wrote something, it will often say yes even if the text was not generated by AI.