Whole Class Gets Fail Grade After Falsely Being Accused Of Cheating With ChatGPT

An unexpected side effect of the artificial intelligence wave? Being gaslit into thinking you’ve used it, even if you haven’t.

According to the Rolling Stone, an entire class of seniors at Texas A&M University–Commerce was accused of taking getting ChatGPT to generate their final essays. The agricultural professor had run their work through the chatbot, and upon getting indication from the AI that the papers might have been plagiarized, he proceeded to flunk them with an ‘X’ grade.

The educator, who also works as a rodeo instructor at the campus and whose name was published in the report, sent out an email to the class explaining why no one had passed. “I copy and paste your responses in [ChatGPT] and [it] will tell me if the program generated the content,” he wrote. To back up his findings, he said he’d even double-checked with the language-processing tool to verify that the papers were plagiarized.

The students had their diplomas withheld due to their alleged misuse of ChatGPT, although some of them had already attended their convocation ceremony earlier this year.

The class was given the option to submit a makeup assignment, or receive a failing grade, which would mean they were not allowed to graduate.

Unfortunately, the professor had been misguided. ChatGPT, however helpful it may be, should never be used as a plagiarism checker. The tool is known to claim all kinds of texts as its own, even classical works. The plagiarism detector released by ChatGPT’s creator OpenAI, too, is prone to giving false signals.

There are some specialized tools that scan bodies of text for AI-manufactured content, like Originality.AI, Copyleaks, and GPTZero, though those have varied accuracy rates too.

When some students tried to prove their innocence by sharing the timestamps of their Google Documents to the professor, he purportedly replied: “I don’t grade AI bulls**t.” At least one of them managed to persuade him and got an apology, though.

Quoting one Redditor, ‘DearKick’, who claimed to be the fiancé of one of the students affected by the unusual grading criteria, the Rolling Stone says the professor had misspelled the bot’s name as “Chat GTP” in his email.

Interestingly, when another Redditor ran the abstract of the professor’s doctoral dissertation through ChatGPT, they got this response: “Yes, the passage you shared could indeed have been generated by a language model like ChatGPT, given the right prompt.”

The university tells PC Magazine that “no students failed the class” as a result of the misunderstanding, and that their graduation status hasn’t been affected.

[via Digital Trends and Rolling Stone, cover illustration 277128883 © Inkdropcreative1 | Dreamstime.com]