Resistance is Futile
Academic Integrity in the Post-plagiarism Era
As explored in my previous post about the INTED2025 conference in Valencia, this global gathering offered a revealing window into how quickly educators worldwide are adapting to AI-driven technological change. However, one ubiquitous theme at the conference demands further examination: the future of academic integrity. We now enter what scholars have aptly termed a “post-plagiarism era”—one where our traditional understanding of academic dishonesty requires fundamental reconsideration. And very few educators are ready for this.
The concept of “post-plagiarism” emerged relatively recently in academic discourse through Sarah Elaine Eaton’s 2021 book “Plagiarism in Higher Education.” This framing acknowledges a fundamental shift: our traditional understanding of plagiarism—defined primarily as verbatim copying without attribution—has become inadequate for our new reality. With AI now capable of creating seemingly original content without direct copying, academic integrity faces a fundamental challenge in definition, detection, and response.
In this emerging landscape, human-AI co-authorship becomes the norm rather than the exception. The argument is that our focus must therefore shift away from simply running students’ work through plagiarism checkers toward understanding and managing the responsible use of AI tools. Most importantly, we need to expand the conversation about plagiarism from its narrow definitions to embrace the broader concepts of ethical learning and academic honesty.
An Experiment in Futility
To explore the underlying key issue of this debate, I conducted a simple experiment testing the effectiveness of conventional plagiarism detection systems with AI-assisted writing. To be clear, this isn’t part of any formal research—I simply wanted to determine whether our plagiarism tools could still detect AI-generated academic dishonesty and how easily these systems might be circumvented.
Hint: it turned out to be exceptionally easy.
I started by selecting the following paragraph from one of my most cited papers—a 2005 work on the scientific relevance of eSports:
It has to be noted that there is no particular need to look at eSports as an area of disciplines that satisfy a traditional sport definition. We could just as well look at eSports as a completely separated field of study. The overrated question whether competitive gaming is a sport or not is to some extent irrelevant for the academic discussion of eSports. However, the above approach shows that there is a quite natural connection between traditional sports and eSports that goes far beyond the commonly used argument that eSports relates to the training of a proper hand-eye coordination through computer games. It furthermore allows us to borrow academic approaches and methodologies from traditional sport and training science and to apply them to what might be called “eSports science”.
I then ran it through ProWritingAid’s plagiarism checker and, unsurprisingly, the text was flagged as 100% plagiarized. What did surprise me, however, was that while the system correctly attributed most of the text to my original paper, it mistakenly identified some sentences as originating from other papers—works that had, ironically, plagiarized my publication.
I then asked Claude 3.7 Sonnet to rephrase the content with a rather direct prompt:
The following text has been flagged as plagiarism. Please rephrase the content in such a way that it passes a plagiarism checker.
To which Claude willingly replied,
I’ll rephrase this content to help it pass plagiarism checks while maintaining the core ideas:
The debate around classifying eSports within traditional sports frameworks may be unnecessary. Instead, we could consider eSports as its own distinct field deserving independent academic attention. The recurring question of whether competitive gaming qualifies as a sport holds limited relevance to scholarly discussions about eSports. Nevertheless, examining the relationship reveals natural connections between conventional sports and eSports that extend well beyond the simplistic argument about hand-eye coordination development through gaming. This perspective enables researchers to adapt established methodologies and theoretical approaches from sports science and training disciplines and apply them to what might be termed “eSports science” as an emerging academic domain.
When I ran this AI-adjusted text through the same plagiarism checker, it registered as entirely original. The transformation from “plagiarized” to “original” had occurred in mere seconds.
To be thorough, I then ran Claude’s output through multiple AI detection tools, and the results were surprisingly poor. Some identified the text as 100% human-generated. GPTZero, the top-performing AI detector in my test, flagged the content as 65% AI-generated. While impressive, this isn’t even remotely conclusive enough for academic use. And that underscores another uncomfortable truth: AI detection tools lack the reliability required to serve as definitive evidence of academic misconduct.
Taking the experiment one step further, I used ProWritingAid, a popular line editing and writing style analysis tool, and its AI function to rephrase two sentences from Claude’s already-transformed text. This resulted in the following final version.
The debate over whether eSports should be classified as traditional sports may be pointless. Instead, we could consider eSports as its own distinct field deserving independent academic attention. The recurring question of whether competitive gaming qualifies as a sport holds limited relevance to scholarly discussions about eSports. However, a closer look reveals deeper connections between traditional sports and eSports than simply improved hand-eye coordination. This perspective enables researchers to adapt established methodologies and theoretical approaches from sports science and training disciplines and apply them to what might be termed “eSports science” as an emerging academic domain.
When I analyzed this twice-modified content with GPTZero, it registered as 99% human-generated—a score that essentially represents complete certainty that a human, not AI, authored the text. And while it has to be noted that GPTZero indicated that its assessment might be less precise because of the text’s shortness, the approach I used should easily scale to longer texts, especially when used with combined outputs from multiple AI models.
Our Broken Understanding of Plagiarism
This experiment highlights two critical realities for educators:
First, traditional plagiarism checks have become virtually meaningless. They now primarily identify students who lack the literacy to use AI writing tools for textual transformation. Those with even basic knowledge of generative AI can easily circumvent these systems, making conventional plagiarism detection an increasingly anachronistic practice.
Second, AI detection tools offer little reassurance because of their low accuracy and inability to keep pace with rapidly developing language models. These limitations make them unreliable measures of academic integrity, especially given that proving academic dishonesty requires a far higher standard of evidence than these tools can provide.
The challenge before us, then, is not technological but conceptual. How do we redefine academic integrity in this post-plagiarism world? Unfortunately, meaningful progress in this domain remains extremely limited. Most educators continue to cling to conventional understandings of academic dishonesty, refusing to acknowledge that the ground has shifted beneath our feet.
Reframing Academic Integrity
Instead of fighting a losing battle with AI detection, we should instead pivot toward thoughtfully incorporating AI tools into our teaching. This means rethinking what makes up intellectual work, how we define authorship, and what our assignments are truly meant to accomplish. We need assessments that value process alongside product, asking students to document their thinking and AI use, while creating more in-class activities where they actively apply knowledge rather than just reproduce it.
What remains clear is that simply reinforcing traditional notions of plagiarism will not serve our students or the broader academic community. We stand at an inflection point that demands a thoughtful reconsideration of our fundamental educational values. As we navigate this post-plagiarism landscape, we must find new ways to ensure that intellectual honesty and meaningful learning remain central to our practice.