(Image credit: Pxhere.com)
As a conduct administrator whose primary duties are adjudicating alleged violations of the Honor Code, understanding how artificial intelligence is being used by college students to aid in the completion of the work has become a primary responsibility. Since the fall semester of 2022 when ChatGPT became more accessible, students have found creative means to use AI to enhance their work or complete it entirely. A common question I receive is “Do you use AI Detectors?”. The answer is yes and no.
The Problem with Detection Software
Facutly and students quickly adopted the use of artificial intelligence detectors such as ChatGPT Zero to use as evidence proving a written assignment was likely generated by artificial intelligence or by a human. When cases started coming to our office and a professor had already told a student they used a detector, and this was the cause for the report, students who wanted to fight the allegation would find other detectors that said the opposite. We would receive reports that documented a back and forth between faculty and students that were arguing more about the validity of the detectors used, as opposed to the original cause for concern that led to a detector being used by the faculty in the first place.
We would review the reports and discuss them with students, but found other methods of investigating by asking questions grounded in how the student described completing an assignment, as opposed to arguing the validity of a report from a random AI detector.
Perhaps the most interesting about reports from ChatGPT Zero was something that their website has posted about using its services in offices like ours. The FAQ section of their website states:
“The nature of AI content is changing constantly. These results should not be used to punish students. We recommend educators use our behind-the-scenes Writing Report for a holistic assessment. See our FAQ for more information…
Our classifier is not trained to identify AI-generated text after it has been heavily modified after generation (although we estimate this is a minority of the uses for AI-generation at the moment).
Currently, our classifier can sometimes flag other machine-generated or highly procedural text as AI-generated, and as such, should be used on more descriptive portions of text.” (GPTZero, 2025)
This uncertainty or disclaimer has made it difficult for our office to use the reports as fact and instead we review them, but as an addition to a report to our office as opposed to the entire report.
Why Did the Faculty suspect AI in the First Place
When I am not adjudicating cases, I am consulting with faculty about academic integrity and the adjudication process. In these conversations over the past 3 years I have had to help faculty find value in their gut reaction that led them to check a student’s paper as AI-generated. This is more important to me as an investigator and is more impactful for a student to hear in a report.
On one hand, faculty at first were feeling some guilt over trusting these detectors in some cases, while others viewed them as fact.
When I speak with faculty, I ask them to walk me through why they chose to check a student’s paper for AI generation. Faculty would say that sources were not paraphrased correctly, formatting was strange, and the paper did not sound like the student or the other students completing the same assignment in the course. These are things that I focus on when speaking with students and I have found it to be disarming. It forces a student to explain their writing process or lack thereof, which is helpful in hearing these cases.
The New Tool in the Investigative Process
Beyond the traditional rapport building, probing questions, and elements of care that are used in investing cases of academic dishonesty, a new method of investigation was revealed to me by a student accused of cheating by using artificial intelligence to complete their paper.
This student had been accused of using artificial intelligence to at least in part enhance their research paper. When I met with the student, they were respectful and discussed their writing process, word choices, and editing process. They wrote the paper using Google Docs and offered to share their original file document with me, so that I could see their version history to prove they wrote the paper. I was happy to review it and we clicked through the version history together. Ultimately, I could see their writing process and there was no indication of behaviors that would align with the use of artificial intelligence. The documents had been worked on over the course of several days. Formatting was consistent throughout the paper and there was no evidence of copying and pasting, which was likely to be the case if a student had been using an AI-powered editor or generator.
Since that case our entire office now requests original file collaborator links from students if available to aid in our investigations. We also highly recommend that faculty require students to complete their work in OneDrive (if on a Microsoft Campus) or Google Docs (if on a Google Campus). This ensures that the documents will have saved version history that cannot be edited, with the added benefit that students are having their work saved in an online database, so it decreases the likelihood of work being lost.
While not every case is aided in this way, as many students do not complete their work on these platforms, preventative measures such as this do more to deter and investigate a student’s work better than an AI-detector. In cases since, I have asked students questions as I would before, but also ask for collaborator links if possible.
This is just one example of how proactive measures can be far more effective than the current alternative of AI detectors.
GPTZero (2025). November. Answers to GPTZero's most common questions. https://gptzero.me/faq
The author's views are their own.
Adrian Fitzgerald Anderson is a higher education professional specializing in academic integrity, student conduct, and the evolving intersection of artificial intelligence and university policy, with extensive experience at institutions like the University of South Carolina and Appalachian State University.
Thank you for being a member of ICAI. Not a member of ICAI yet? Check out the benefits of membership and consider joining us by visiting our membership page. Be part of something great!