Plagiarism in the age of AI.
The rapidly evolving discussion on plagiarism in the classroom brought on by the proliferation of generative AI: what's actually happening, how to cite it, and why detectors aren't the answer.
Many educators feel uncertain of how students can use generative AI tools to both intentionally and unintentionally plagiarize, heightening classroom mistrust.
Many official citation styles, including APA, MLA, and Chicago, have released comprehensive resources detailing how to cite AI in student work.
AI plagiarism detectors are prone to both false positive and false negative results, can be easily evaded, and have proven to be biased against non-native English speakers.
How students use AI to plagiarize
Generative AI can be used for academic dishonesty in several ways, complicated by the ambiguity in how institutions define plagiarism. Three examples from industry leaders:
Making up sources
Generative AI has been shown to fabricate titles and links to sources, as well as misattribute data to a source.
“ChatGPT will completely make up citations to 'support' the text it generates. Here, in our library, we've already encountered multiple instances of faculty and students looking for books and articles that don't exist.”
— Arkansas State University
Passing AI-generated content off as student work
Without proper citation, the duplication of content originally produced by ChatGPT or other generative AI tools is often considered plagiarism.
“This sentiment was echoed by ChatGPT itself, as reported by the University of South Florida's library.”
— University of South Florida
Failing to credit copyrighted material
It is possible that AI tools reproduce content in their outputs that, if AI were not being used, would need to be cited by students.
“Generative AI derives its output from previously created texts from other sources that the models were trained on, yet doesn't cite sources.”
— St. Edward's University
AI citation
Should student use of AI be cited? If so, how? These questions open a new, confusing aspect of assessment, but internationally recognized organizations have published guidance.
The efficacy of AI writing detectors
Discomfort with student AI use has led to reliance on detector tools that advertise themselves as foolproof. They aren't, and false accusations deteriorate the classroom trust they were meant to protect.
An article from Ars Technica reports that OpenAI itself has repeatedly denounced the efficacy of AI detector tools, including OpenAI's own discontinued tool.
A 2023 Stanford University study demonstrates that AI detection is consistently biased against non-native English speakers, furthering inequalities within academic settings.
Many more detectors have appeared, with continued low accuracy. In response, “AI humanizers” modify AI text so detectors no longer recognize it by varying sentence length and word choice. Turnitin has since released an “AI bypasser detector,” though its effectiveness is unconfirmed.
It's natural to feel apprehensive. Shift the attention toward student-driven work, in-class discussion about AI use, and syllabi built for the modern classroom.
AI plagiarism buzz
Articles from news outlets and industry professionals on the ongoing discussion.
AI plagiarism myths
Six claims we hear in every workshop. Tap one.
Courses & webinars
Short recorded webinars on this topic, and the workshop we run for whole faculties.
One hour to a full day, in person or virtual.
