Part I: “The Death of Plagiarism”
One might read our title – “The Death of Plagiarism” – and think, “Where has this guy been? Hasn’t he heard the rattle and prattle about Harvard’s president plagiarizing parts of (paragraphs of!!!) her doctoral dissertation? It’s all over the news!”
As it should be, but one should remember that Harvard’s president committed plagiarism long before the advent of artificial intelligence (AI), so please, defer your skepticism that “plagiarism is dead” until you finish this short essay.
Some explanation is in order about what triggered me to write this.
Recently I randomly interviewed about a dozen college students. What prompted those interviews was a new experience for me but one now that is almost universally problematic in academia – fraudulent essays generated by artificial intelligence. So, here’s the question I posed to them:
“What percentage of students at this college use artificial intelligence to do written assignments?”
The students’ uniform answer – “100%”
That’s the new reality, at least the new reality that academic institutions and teachers are trying to figure out, especially teachers who require advanced writing assignments . . . with great frustration I might add.
I also interviewed about a dozen college professors and several private-
school teachers and asked this question:
“What percentage of your students are using artificial intelligence to do written assignments?”
The teachers’ answers: “from 60-100% of students use AI to (try and) cheat.”
I’m past forty years teaching, and I don’t think I ever missed a single case of plagiarism. That may sound somewhat braggadocios, but I’m not an exception; any good teacher will always catch plagiarism. It’s easy to detect. At least that’s the way it used to be . . . before artificial intelligence. Sure, there might be that rare exception of a uniquely gifted writer who “knocks my socks off,” but I’ve taught only a few of those. Besides, the method I’ve used over the years has been to require students to handwrite a first draft in my classroom under my immediate supervision (preferably in cursive). That way I “X-ray” precisely what students really know and what they really can do. I could care less if they know how to use crutches like Spellcheck, Grammarcheck, Grammarly, or Chatbox; I want to know what the students know. If I detect plagiarism, I just have to Google it and I’ll find it, as they say . . . Every. Single. Time. Or, if for some reason I couldn’t require students to handwrite a first draft under supervision, “Turnitin” or some other plagiarism detector would easily identify plagiarism for me. At least that’s the way it used to be before Artificial Intelligence. Now, plagiarism is dead.
With the advent of AI, students no longer need to plagiarize; they can dictate an assignment to an AI program like Chatbox and, “Voila!,” out comes a grammatically perfect, well organized essay with above average content. But AI-generated essays are also easy to detect, especially if you’ve gathered previous handwriting samples. But even if you don’t have handwritten first drafts, AI-speak is still easy to detect; the problem is that, if you Google an AI-generated document – if you copy, cut, paste, and search for plagiarism – you get . . . NOTHING! Why? Because, even though the AI-document is fraudulent, it’s also “original” because an AI program composed it. In other words, who needs plagiarism? If an AI-generated essay is undiscoverable through normal means of detecting plagiarism, why plagiarize? Just use Chatbox! So here’s the problem for educators: How can we
possibly detect AI fraud?
That’s also easy; as we’ve already mentioned, a good teacher will always recognize it. As far as proving it, though, especially when previous handwritten samples have not been collected, what’s a teacher to do? One trick is to highlight more sophisticated terminology in a suspicious essay and then set up a meeting with the student to ask him or her to explain complex terms they used and how those terms relate to their essay content. Typically, students will not know the more sophisticated language they used in their AI-generated papers, much less how those terms relate to their essays. I’ve done that several times – quizzed students about the vocabulary in their essays – the most recent time hearing a student respond, “I don’t know those words.” Of course, he didn’t know those words; they were not his words.
More telling are AI-detection programs that can analyze a document for human versus AI origin. It’s disturbing how many essays I’ve submitted to an AI detector that came back as “0% Human.” Of course, eventually, AI will find a sneaky way to evade AI-detection devices. A good teacher, however, will always know more than Grammarly, Spellcheck, Grammarcheck, or an AI detector, which brings us to this foundational and essential principle of academic integrity and excellence – “trust.”
Stated another way, “When it comes to detecting academic fraud, whether via AI, plagiarism, or some other nefarious means, the good teacher is always right.” Of course, there are exceptions to that proverb, but during my long career as both a teacher and administrator, I’ve never seen a single instance in which a teacher was wrong about cheating; when a teacher smells a rat, there’s a rat. This means that, unless AI detection keeps up with AI deception, schools, colleges, and universities will either have to do what they are doing now – concede defeat to AI – or increasingly empower their teachers to say, “I can’t accept this; this is not your writing.”
The Pandora’s Box of digital “learning” is already open and will never be closed, no doubt expanding into larger and stranger technologies than we can possibly imagine, which brings me back to AI detection:
Make the qualified teacher, not technology,
the final arbiter and judge of originality.
They’re intelligence is not artificial.
You can trust them.
About the author . . .
Dr. Hal Brunson is the consultant for Paris Classical Academy, which plans to open its doors in August, 2024. An active writer, he holds the doctorate from the University of Texas at Dallas and has taught at the undergraduate and graduate levels, as well as served as consultant and administrator for numerous classical-Christian schools.
(In our next segment, we will explore how Classical, Christian Education represents an insurmountable obstacle to AI fraud.)