An Animated Cat Singing The iPhone Ringtone Became A Meme
In March, South Park released an episode skewering the mania around AI generative text creator ChatGPT by having the fourth-grade boys use the program to pretend to be caring boyfriends and also write their essays. When their teacher, Mr. Garrison, finds out that there's a program that can write essays for the students, rather than get suspicious, he has the bright idea to use ChatGPT to save him the work of grading papers.
A bizarrely close recreation of the episode apparently went down at Texas A&M-Commerce when a professor used ChatGPT to grade their student's papers. However, unlike Mr. Garrison, the professor basically flunked the class and thus denied the students their diplomas on the suspicion they'd all used ChatGPT.
According to a report from Rolling Stone looking into a Reddit post by DearKick, Dr. Jared Mumm failed an entire class after deducing they'd all supposedly used ChatGPT to complete their final three assignments for the semester.
Mumm emailed the students afterward explaining how he'd made the deduction and offered them the chance for a makeup assignment.
According to his email, Dr. Mumm had himself used ChatGPT like Turnitin, a program that will scan student papers for plagiarism. Mumm asked ChatGPT if it had written the assignments his students turned in, and ChatGPT replied that it had.
The only problem is that ChatGPT doesn't work like Turnitin (or an actual program intended to detect if writing is AI-generated, like ZeroGPT) and is not smart enough to deduce if something has been written by AI. According to Rolling Stone's Miles Klee, the program will claim it wrote Crime and Punishment with "very little prodding."
In proving the folly of Mumm's AI-detection method, Redditors asked ChatGPT if it had written both Mumm's email to students and a section of his dissertation, to which it replied that it had.
DearKick claims to be the fiancé of the student who shared the email, and says their partner was baffled because they'd never heard of ChatGPT before. Other students attempted to prove to Mumm that their papers were genuinely their own thoughts and words, but purportedly received responses from Mumm saying, "I don't grade AI [BS]."
After the story was published, Texas A&M-Commerce representatives emailed Rolling Stone and denied that anyone had received a failing grade and added that it would be investigating how to best help professors detect AI in the future.
"Currently, several students have been exonerated and their grades have been issued, while one student has come forward admitting his use of ChatGPT in the course," the statement read. "Several other students have opted to complete a new writing assignment made available to them by Dr. Mumm."
Display Comments