Valpo’s Honor Code and AI

By Cynthia Rutz, Director of Faculty Development, CITAL

Faculty are increasingly encountering AI in the classroom, whether they want it there or not. Not surprising, then, that there is an increase in the number of Honor Council (HC) cases involving the unauthorized use of AI. In this article, the Honor Council Executive Board fills us in on how AI is impacting Honor Council cases and how faculty can help their students to use AI responsibly.

I asked the HC Board what kind of cases they have seen around AI. They cited three examples:

• Cases where a faculty member thinks a piece of student writing was written by AI because the style is very different from previous writing.
• Cases in which a faculty member finds fake sources in a research paper.  AI is known for “hallucinating” sources, i.e., citing academic papers that do not exist. (Although this may be changing, see this article: AI Hallucinations May Soon Be History.)
• In one case, a supposed interview paper seemed suspect because the organization that the interviewee was from was found not to exist.

 In cases such as these, faculty sometimes provide evidence from one of the many AI-detection tools out there. However, because these tools are notoriously unreliable, they can never be the only evidence provided (AI Detectors Don’t Work. Here’s What to Do Instead).  

Based on these AI cases, the HC Executive Board had the following advice for faculty:

1. Whenever you assign a paper that requires sources and citations, be sure to check the sources. Some of the hallucinated sources mention real writers and/or real journals in the field; however, the citation as a whole can still be made up by AI.

2. Even if you think you have only limited evidence that something is written by AI, be sure to provide all that evidence before the hearing, including prior pieces written by the student. The hearing panel will take into account the difference in writing style and ask the student why their writing has changed.

3. Consider making it a policy that your students must write their paper in Google Docs and share it with you. Google Docs has a version history, so you will be able to see how the writing has evolved over time. If instead of slow improvement you see that suddenly a fully written paper appears, you can cite that as evidence of use of AI.

The HC Executive Board also had some ideas for how faculty can help students to avoid students misusing AI in the first place. Here are some of their ideas:

1. Always explain WHY, whether you are authorizing AI in certain instances or forbidding it altogether. For example, for a writing assignment, explain the importance of the process rather than the product (and assign process points accordingly). You can tell them that AI is not authorized for certain parts of the writing process because they are learning communication skills that they will need in their future careers. 

2. It can be a problem when some classes integrate AI too early in the writing process, before students have learned how to write on their own. This can lead to students using AI as a crutch and then not growing their writing skills throughout their college careers. So consider barring AI from early writing classes.  Let students know that you do so because you want them to develop their own voice first, since that is what sets them apart.

3. Be crystal clear in the syllabus about when AI use is authorized and when it is not. It could be useful to use graphics, such as a red stoplight for parts of the assignment where AI is not permitted and a green stoplight where it is allowed.

4. For each and every assignment description, be equally clear about which parts of the assignment include the use of AI and which parts forbid it.

5. One faculty member had a very clear AI statement both in their syllabus and on their assignments about what constituted authorized and unauthorized use of AI. Although there were several HC cases in that class, those cases were adjudicated quickly because it was quite clear to the HC panel when students were or were not violating authorized use.

6. Do not assume that you know what AI-written text looks like. Instead, try running one of your assignment prompts through several AI programs–ChatGPT, Claude, Gemini–so you can see what it reads like. You could also give these AI versions to your students to critique. Or you could change your prompt if you find that AI can do it too easily.  

7. Many students do not consider themselves good writers and know that they need help. To avoid having them go to AI for that help, consider promoting campus resources such as the writing center, where they can get help from fellow students. Lower the barrier to seeking help by making a writing center session mandatory for one assignment or by giving extra-credit points for using the writing center.  You could also hold a class help session in the writing center run by the student writing consultants.

8. AI is here to stay, and it is getting better and better all the time and harder to detect, especially for essays. So consider giving alternative assignments, including in-class assignments, that cannot be done by AI. 

9. Keep in mind how fast AI is evolving. As soon as a faculty member or student figures out how to work with it, the technology evolves again. So be willing to be a continuous learner with AI. 

Many thanks to the HC Executive Board members who provided information for this article: Executive Chair Ana Bozinovski (student), Alison Downey (Library), Jiun-Lin (Alex) Chen (Business), and Jon Beagley(Math & Stats).  The HC Executive Board has also put together a useful guide for faculty, “What Faculty Need to Know About AI,” that is located at the bottom of the first page of the Honor Code website here:  The Honor Code.