Professor Catches Students Plagiarizing With AI

Blog, AI Detector

Professor Catches Students Plagiarizing With AI

In a situation so ironic and meta, it’s almost hard to believe, Megan Fritts, an assistant philosophy professor at the University of Arkansas, Little Rock, caught numerous students plagiarizing with AI on their first assignment for her Ethics and Technology class.

AI Plagiarism in an Ethics and Technology Class

The assignment was simple: “Briefly introduce yourself and say what you're hoping to get out of this class”. The purpose of the assignment was two-fold: to learn about her students and to get them acquainted with Blackboard’s discussion board feature.

Having used ChatGPT to generate their essays, many of the students submitted similar outputs that instead of stating their desired expectations of the class, regurgitated descriptions of what an ethics and technology class is. Such personalized requests are often too complex for ChatGPT’s algorithm, so it comes as no surprise that Fritts was able to spot the watermarks of AI writing right away.

"When you're a professor, and you've read dozens and dozens of AI essays, you can just tell," Fritts responded when asked how she was able to detect the presence of AI.

The students, perpetuating their interest in understanding ethics and technology, did the right thing and immediately confessed to using AI to write their introductory essays.

"They all owned up to it, to their credit, but it was just really surprising to me that — what was supposed to be a kind of freebie in terms of assignments — even that they felt compelled to generate with an large language model."

The casualness of students using AI to write an introductory essay speaks to two phenomenons: First, that students are using AI writing tools for school work at astounding rates.

Where it was reported that 86% of college students are using AI in their studies.

Second, that students don’t see the need to put effort into meaningless work. In the replies to Fritt’s now viral X post detailing her story, many of the replies argued that of course students would be using AI for assignments they deemed “busywork”, comparing the use of AI to write essays to using a calculator to solve math equations.

Fritts disputes this comparison, especially in the case of humanities assignments, which she feels are meant to shape people not dispense content.

"The goal is to create liberated minds — liberated people — and offloading the thinking onto a machine, by definition, doesn't achieve that," she said.

Although using the word “liberated” implies a political context or perhaps agenda, Fritts does touch upon the academic consensus that doing school work is essential to developing thinking ability. Whether by using generative AI or simply by harboring a technology addiction, our thinking ability from our decision making, critical thinking faculties, or even our attention spans, are compromised.

Tech Solutions for Tech Problems

There are numerous instances of students facing severe consequences for using AI, but it's teachers who are suffering with the burden of finding new ways to make meaningful learning experiences that counteract AI usage.

Many are feeling helpless, with so many students using AI writing tools, whether undetectable or not, and with the outright failure of AI detection tools to properly do their jobs, the future of academia seems dim to some.

Even Fritts herself acknowledges that the burden of curbing the AI plagiarism problem can’t be on the students. Instead, institutions are realizing there is no way out other than embracing AI in academia.

Higher Education institutions have been utilizing AI for a variety of tasks from AI detection software, to generative AI for curricula, syllabi, and homework.

Whatever faculties students don’t develop from doing old-fashioned school work, might be developed from learning AI literacy and applying the technology to complicated problems.

The Future of AI in Academia

As detailed in this Forbes article predicting AI trends in academia, one should expect the institution to buckle against technological change by restructuring the way students take on their writing assignments. Many of these assignments will be done in class to curb AI cheating, from brainstorming to submission, students will write their essays in class.

Once the systems start implementing more customization and personalization, they will be able to shape learning experiences to each individual student in a classroom, so no student gets left behind.

Should The Use of AI Tools be Considered Plagiarism?

Using AI-generated content isn’t traditional plagiarism. Plagiarism is the theft of someone else’s words or ideas without proper credit. Meanwhile, AI-generated text is always original. Catch students using artificial intelligence, they will be accused of academic dishonesty under the category of “AI plagiarism”.

So no, using AI writing is not plagiarism. You can go as far as calling these students cheaters, but that might not age well with the way things are going.

AI Tools for Submission and Detection

Had Fritts’ students used undetectable AI writing tools, perhaps this story wouldn’t have ever been written. Unlike OpenAI’s ChatGPT or other AI chatbots, undetectable AI writing tools like StealthGPT don’t feature the watermarks of AI-generated text and instead replace them with a human writing style.

For instance, StealthGPT bypasses AI detectors like Turnitin, GPTZero, Originality.ai, and more. These AI checkers usually have high false positive rates, like Turnitin which acknowledge a 4% false positive rate at the sentence level.

To understand why this happens so often, you need to understand what is an AI detector. An AI detector analyzes text for AI writing and that just means it tries to find word choice, sentence structure, and sentence length that are so simple and precise, they were probably written by an algorithm that optimizes for utility. It’s rather easy to find student writing that uses simple and precise language, which was the case when Stanford discovered bias in AI detectors against students that didn’t speak English as their first language.

Written By

Rob Shepyer

Rob Shepyer

Undetectable AI, The Ultimate AI Bypasser & Humanizer

Humanize your AI-written essays, papers, and content with the only AI rephraser that beats Turnitin.