Several seniors at Texas A&M University-Commerce, who had already completed their graduation ceremony this year, faced a temporary withholding of their diplomas due to a professor's inept use of AI software in checking their final assignments for supposed "cheating."
From Rolling Stone:
Dr. Jared Mumm, a campus rodeo instructor who also teaches agricultural classes, sent an email on Monday to a group of students informing them that he had submitted grades for their last three essay assignments of the semester. Everyone would be receiving an “X” in the course, Mumm explained, because he had used “Chat GTP” (the OpenAI chatbot is actually called “ChatGPT”) to test whether they’d used the software to write the papers — and the bot claimed to have authored every single one.
“I copy and paste your responses in [ChatGPT] and [it] will tell me if the program generated the content,” he wrote, saying he had tested each paper twice. He offered the class a makeup assignment to avoid the failing grade — which could otherwise, in theory, threaten their graduation status.
There’s just one problem: ChatGPT doesn’t work that way. The bot isn’t made to detect material composed by AI — or even material produced by itself — and is known to sometimes emit damaging misinformation. With very little prodding, ChatGPT will even claim to have written passages from famous novels such as Crime and Punishment. Educators can choose among a wide variety of effective AI and plagiarism detection tools to assess whether students have completed assignments themselves, including Winston AI and Content at Scale; ChatGPT is not among them. And OpenAI’s own tool for determining whether a text was written by a bot has been judged “not very accurate” by a digital marketing agency that recommends tech resources to businesses.
But all that would apparently be news to Mumm, who appeared so out of his depth as to incorrectly name the software he was misusing. Students claim they supplied him with proof they hadn’t used ChatGPT — exonerating timestamps on the Google Documents they used to complete the homework — but that he initially ignored this, commenting in the school’s grading software system, “I don’t grade AI bullshit.” (Mumm did not return Rolling Stone‘s request for comment.)
Oh my goodness...
I'm unsure about the age of this professor, but they appear to be someone who recently discovered ChatGPT and are unfamiliar with it. Ironically, they believed they were being innovative by using AI software to identify cheating.
Unfortunately, their plan backfired completely.
