Recently, I caught up with a friend who started a new job as a teacher’s aide (TA) at a local university. As a TA, one of the responsibilities is to review and grade student papers. My friend shared with me that AI is used in contrasting ways at the university, which resulted in a protracted conversation given my interest in the topic!

For students, some of them are taking advantage of ChatGPT and similar technologies to help write part of their papers to minimize the amount of research and writing. This is understandable given the workload a student is expected to do to attain their degrees and diplomas. What the students do not understand is that GPT technologies can be detected by teachers and TAs. Ironically, these students are working towards a technology related degree!

My friend has an advanced academic background that includes linguistics in multiple languages. I am told that teachers and TAs can detect when students augment their original thoughts and expressions using technologies like GPT. One of the ways in which this is noticed is when writing styles and the “voice” of the author suddenly changes. While GPT can help students who struggle with written expression of their original thoughts, GPT is not yet advanced enough to consider the writing styles or voice of the author. This is not good for the reputation of the student in the eyes of teachers and TAs.

However, my friend was very quick to highlight that a student’s potential use of AI technology has already surfaced several false positives. No specifics were provided.

While the ethics of using AI and GPT technology by students can be debated ad nauseum, I am neither qualified nor am I willing to engage in such a debate.

But students are not the only ones to take advantage of AI technology. Universities are using AI in a different way. They use AI to detect plagiarism. Plagiarism is possibly the most egregious offence in academia and can carry penalties that range from severe disciplinary action to immediate expulsion.

In this use case, I find the use of AI for this purpose to be practical, effective, and time efficient. AI technology is used to comb through the universe of digitized published work that is accessible to the university to detect if a student is passing off someone else’s original work as their own.

A student’s reference to or quotation of another author’s work does not equate to the “P” word so long as the student properly cites and attributes the referenced section to the original author. In fact, in many cases, it is encouraged as it helps students to defend their conclusions. (I use the term “P” word given the gravity of even the hint of such accusation).

I am told that this use case of AI has already proven itself to be quite effective. The university parses a student’s work through an AI application that searches the body of work of that relates directly or is adjacent to the topic at hand. This body of work includes the university’s own archive of papers submitted by previous students.

In one case, apparently, the AI application quickly detected the possible transgression by one student. (I mention possible, because this possibility is still under review). This student is accused of copying and pasting several pages (yes, several) of a previous student’s work. Apparently, the accused offence was quite blatant – the student even copied the spelling mistakes contained in the previous student’s paper. As the cliché goes, “oh, no you di’int!”

For a university, the use of AI for use cases such as this is obvious. Before the assistance of AI technologies, the detection of the “P” word, even against papers submitted by previous students archived within the same university, may not be easily recognized. Absent the use of AI, even the most knowledgeable and experienced peer reviewer or dissertation jury would need to have an encyclopædic knowledge of all papers ever published on a particular topic and adjacent/complementary topics to suggest that there has been misappropriation of another author’s work.

I recognize that some people struggle with the written expression of their genuinely original ideas. In these cases, AI technologies like GPT can and does help. In my opinion, AI is still in its infancy. Reliance on it can help or hinder. In the examples above, AI helps universities, while AI (in the form of GPT technology) can potentially hinder or hurt students.