Photo: Getty Images
A college professor at Texas A&M reportedly failed half of his class after accusing them of cheating. Reddit user DearKick shared a screenshot of an email their fiancee received from their professor, accusing them of using OpenAI's ChatGPT to write their papers.
"In grading your last three assignments I have opened my own account for Chat GTP [SIC]. I copy and paste your responses in this account and Chat GTP will tell me if the program generated the content," Dr. Jared Mumm wrote.
He then informed them that the AI claimed it wrote their papers and gave them a zero.
"I put everyone's last three assignments through two separate times and if they were both claimed by Chat GTP you received a 0."
He offered to let students rewrite the papers, but many claimed they did not cheat and even showed time-stamped proof of their work in Google Docs.
In a follow-up post, DearKick explained that half of the class had been accused of cheating and that their diplomas were on hold while the school investigated the claims.
According to Business Insider, ChatGPT isn't capable of detecting whether or a paper was written by artificial intelligence. While there are some programs that can do that, Dr. Mumm did not use them.
PC Magazine decided to test ChatGPT and inputted articles snippets written by its editors and asked the AI chatbot if they were written by a human or an AI.
"ChatGPT erroneously said it had written the passage. In addition, the AI program has limited memory, along with no capability to recall conversation histories from other users," Michael Kan of PCMag.com explained.
In a statement to PCMag.com, Texas A&M said that none of the students failed the class. While their diplomas were still being held, they were all allowed to graduate with their classmates.
"No students failed the class or were barred from graduating as a result of this issue," the school said. "The professor is working with the students to determine whether AI was used to write their assignments and, if so, at what level."
OpenAi is aware of concerns that its chatbot can be used by students to cheat on their assignments and is working on a solution that would allow the chatbot to detect AI-generated text.