From Yahoo News:
A University of North Georgia (UNG) student is on academic probation after she says she used Grammarly to proofread a paper. The school says the action was taken because they detected the use of artificial intelligence in violation of their plagiarism clause in the student code of conduct handbook.
“It’s just been very frustrating,” UNG junior Marley Stevens said.
Stevens, after submitting a paper for her criminal justice class in October, says she was surprised to learn her professor gave her a “0” for the assignment and reported her to the Office of Student Integrity.
“He was like you used AI on your paper, you get a zero, that was it,” Stevens said.
“I had Grammarly installed in my web browser, but I’ve only ever had the free version, so all it did was fix my punctuation and my spelling,” she added.
. . . .
She submitted the paper through the program Turnitin, which flagged it for the use of AI.
Turnitin launched an AI writing detection feature in March 2023 to find when AI writing tools generate words used in submissions rather than the students’ own writing.
Earlier this month, Stevens learned she’d been placed on academic probation.
Grammarly says its suggestions for grammar and spelling changes aren’t made through generative AI, which is an algorithm which can create new content on its own.
Grammarly sent FOX 5 a statement reading in part:
“Grammarly’s trusted writing support helps students improve their writing skills by offering suggestions for spelling, grammatical correctness, clarity, concision, and tone. These suggestions are not powered by generative AI and can still be accessed even when generative AI features are deactivated or not used by the student. However, some third-party tools may mistakenly identify any use of Grammarly as generative AI. We encourage institutions to establish clear policies on acceptable AI usage and adhere to those guidelines when assessing student success.”
Stevens said she’s used Grammarly on other assignments before without problems.
“I had teachers before who made us install it and turn a screenshot in that we had done so-and-so I’ve written my papers the same exact way all through college in a Google Doc with my Grammarly extension. I’ve never had any problems,” she explained.
. . . .
Regarding its AI policies, the University of North Georgia issued a statement reading in part:
“Our faculty members communicate specific guidelines regarding the use of AI for various classes, and those guidelines are included in the class syllabi. The inappropriate use of AI is also addressed in our Student Code of Conduct.”
Stevens took to TikTok to share her story, which has millions of views.
Stevens’ academic probation currently lasts until February 2025.
Link to the rest at Yahoo News and thanks to F. for the tip.
PG wonders if the professor in the OP actually read the paper in question or simply relied on one computer program accurately detecting the use of another computer program.
One wonders just how enthusiastically (undisclosed, naturally) services like Turnitin are using precisely the same premises as LLM-based systems… oh, wait, it’s fully disclosed, just not using the same terminology.
So we have an AI-based-under-different-naming-conventions system accusing something else of using AI. With no human supervision of either. The irony of that is a bit much.
One also wonders if this professor had ever heard of the inability to copyright material whose creative expression is constrained by the limited number of ways facts (and analysis within the constraints of narrow disciplines) can be expressed. And since that’s exactly what a student paper is/is supposed to be — especially in the kind of course being taken by sophomores — maybe this professor had inappropriate expectations in the first place.
Of course, in law school the sort of thing this student was accused of would be encouraged: The mark of a “great lawyer” is the ability to prove a dubious proposition by “proving” that some authority has said exactly the same thing before. That would definitely set off any LLM-based system’s detectors, because it’s just a human running the LLM instead of a computer running the LLM.
Humans have several LLMs in their heads: education, experience, news, biases, rumors, urban legends…
Very litle of which gets challenged.
Experience is not an LLM. Most of it has very little to do with language of any kind.
Given the post below this one, where turnitin operates on words or patterns of words its probably a false positive anyway. Checking spelling and grammar only should not change the actual words