Over the summer, it finally happened to me: while reading a student essay in an introductory online college course at one of the universities where I teach, I began to suspect not only that this student had not written it but that no mind human had written it. . In other words, it was the work of ai (artificial intelligence).
Like many educators around the world, I was thrown into a brave new world of modern teaching for which I had not been trained and for which an adequate response was unclear.
Although I was aware of many effective ai detection tools I had also heard horror stories of false positives from these same tools. My institution had recently announced a policy prohibiting professors from accusing students of using ai in their work. This gave me an institutional framework in which to operate, but I also did not want to reward a student with a positive grade for using a typewriter for him.
This is what I did, both in this case and in subsequent cases of suspected ai use that I encountered.
<h2 id="1-detecting-ai-i-trusted-my-instincts-to-a-point”>1. ai Detection: I trusted my instincts to some extent
For me, the first step in this process was to use common sense when reading student work. ai tools have gotten good at writing in some ways, but they tend to be bland: the written equivalent of white noise or smooth jazz.
Additionally, the student I first suspected submitted a paper that was markedly different from any he had submitted earlier that quarter. But as good as I think my spidey sense is at detecting ai writing, it definitely wouldn’t hold up in a court of law, so I knew I needed more proof before continuing.
<h2 id="2-i-trusted-ai-detectors-to-a-point”>2. I trusted ai detectors to some extent
My next step was to run part of the suspicious ai-generated document through an ai detection tool. The paper found that more than 95 percent of the probabilities had been generated by ai in two different readings of detection tools. As a control, I ran parts of two other student papers that appeared to be written by humans through the same detection tools; these turned out to be human-generated, as I suspected.
Advice: If you don’t have a favorite yet, check out our ai Detection Tools Guide. Also, remember that there have been ai-detectors-over-fears-students-falsely-accused-cheating-2023-9″ target=”_blank” data-url=”https://www.businessinsider.com/universities-ditch-ai-detectors-over-fears-students-falsely-accused-cheating-2023-9″>Privacy concerns are raised about ai detection tools., so be sure to never include a student’s identifiable information through one of the tools. Finally, please note that ai-detectors-over-fears-students-falsely-accused-cheating-2023-9″ target=”_blank” data-url=”https://www.businessinsider.com/universities-ditch-ai-detectors-over-fears-students-falsely-accused-cheating-2023-9″>Some universities have stopped using ai detection tools. due to concerns about accuracy.
3. I talked to colleagues
After the positive readings from the ai detection tools, I was almost certain that the article had been generated by ai. However, he was aware that technology/college-students-cheating-software-honorlock.html” target=”_blank” data-url=”https://www.nytimes.com/2022/05/27/technology/college-students-cheating-software-honorlock.html”>false positives have occurred and students deserve the presumption of innocence when it comes to academic violations.
I contacted a colleague in my department who supervises my work and informed him of the situation and what I had found. He read the document and ran it through a separate ai detection tool, and agreed that it was almost certainly generated by ai. He then called a superior of his to get a third opinion and discuss a course of action.
This process helped protect the student by ensuring that they were not misinterpreting the situation in some way. Talking to others also protects me as a non-tenured educator because it documents that I am acting in accordance with the university’s ai policies. This is especially important because in many institutions these policies are constantly evolving and in some cases have not yet been formalized.
<h2 id="4-i-spoke-with-the-student-without-accusing-them-of-using-ai-xa0″>4. I spoke to the student without accusing him of using ai
My colleagues and I decided that the best course of action would be to try to talk to the student. During this conversation, I was asking them how they conducted their research for the article without mentioning ai or accusing them of anything. I reached out via email and asked to speak to the student by phone or video call. When the student did not respond, I emailed him specific questions about the content of the work and told him that I needed answers before I could grade his work.
At the same time, I sent emails to the entire class informing them that I had encountered suspected ai use in the classroom. These announcements reminded students that the use of ai essay writing was against course policies.
In the end, the student never responded and the work received a zero. Since then, I have come across several other cases of suspected ai use and they all played out more or less the same way. None of my students have been disciplined for violating academic standards, but none received credit for work they did not complete, which is my main concern.
It is a small victory, but I think it is important, for human writing.
