Professor accuses student of forging essay with ChatGPT

Professor at Furman University (the United States), Darren Hick, said that his student used the ChatGPT neural network to forge the essay.

According to the professor, one of his students wrote a paper on David Hume’s paradox of horror, but the text contained numerous errors. Hick suspected fraud by the student and checked the essay with another tool.

“The first sign that I’m dealing with AI is the text’s meaninglessness. In the essay, the views on Hume’s paradox of horror are described confidently and flawlessly, yet they were all completely wrong,” Hick wrote.

It turned out she used ChatGPT to write the paper:

“Any arbitrary chunk of this text [the tool] marked as fake with 99.9% accuracy, while the works of other students proved to be 99.9% real.”

Hick also noted some structural features of the texts that betray the “handwriting” of ChatGPT. According to him, such papers are easy to distinguish even at a glance. However he believes that in the future detecting fakes will become more difficult.

“In my case the student admitted using ChatGPT, but if she hadn’t done so, I probably wouldn’t have enough evidence. It’s a very new [technology],” the professor said.

He urged educational institutions to immediately develop standards to combat this kind of fraud.

In December, OpenAI introduced the chatbot with the ChatGPT AI.

In October, Reddit users shared stories of how they applied language models to write scientific papers.

In July, GPT-3 wrote a scientific paper about itself.

Subscribe to ForkLog news on Telegram:ForkLog AI — all the news from the world of AI!