‘CheatGPT’: new AI platform poses a threat to academia

‘CheatGPT’: new AI platform poses a threat to academia

When Associate Professor of Politics Khalil Habib saw the writing of artificial intelligence tool ChatGPT, he removed all his essay assignments for his courses.

“I decided to do in-class exams this semester until I get more of a handle on this new technology and how to work around it,” Habib said.

Available for free online, this new program can write an essay, explain the Pythagorean Theorem, and even write computer code. ChatGPT is a thrill to use at first, but professors across the country, including Habib, have become worried about its implications for academic dishonesty.

“I think it has the potential to be bad if students use it to write papers,” Habib said. “Learning to write is learning how to think and articulate ideas and arguments. Shortcuts or attempts to circumvent this process are terrible on many levels.”

Hillsdale College Provost Christopher VanOrman said Hillsdale College has been watching ChatGPT closely.

“In the last few months, we’ve had several conversations with IT and with a group of our faculty on how best to address this issue,” VanOrman said. “We will continue to counteract any adverse effects ChatGPT might have on our campus, but we also know that we have to rely on our trust in our students to abide by the honor code.”

Developers at artificial intelligence company OpenAI, creator of ChatGPT, built a large language model designed to predict the next word of a sentence, and trained it on more than 300 billion words of information, according to Science Focus. The data set included books, articles, and Wikipedia.

If you ask ChatGPT why the Roman Empire fell, you’ll get an answer that could probably pass for a short-answer response in most Hillsdale classes.

“Some factors that contributed to the decline and eventual collapse of the Roman Empire include: economic decline and overspending on military conquests, political instability and a decline in moral values, as well as pressure from foreign invaders such as the Germanic tribes and the Goths,” responds the robot. “Additionally, the empire’s vast size made it difficult to govern effectively and a shortage of military personnel weakened the empire’s ability to defend itself from invaders.”

When a robot becomes too powerful, sometimes the answer is to build a robot to beat it. That’s what Princeton senior Edward Tian did when he built GPTZero.

“I spent New Year’s building GPTZero — an app that can quickly and efficiently detect whether an essay is ChatGPT or human written,” Tian said in a tweet.

When a user copy and pastes text into GPTZero, the program evaluates how formulaic the text seems, and makes an assessment as to whether a body of text was written by ChatGPT.

Tian has continued to develop the program since he released it online Jan. 2. While its efficacy is still being tested, Tian said in a tweet that his newest version, GPTZeroX, is made for educators. The software received 2.2 million requests during its first day online.

Habib said he would consider assigning essays again if there were a bot that could detect ChatGPT writing.

Senior Meera Baldwin said the Hillsdale student body does not seem like it would want to depend on artificial intelligence.

“Hillsdale students tend to trust their own intellect over that of a robot,” Baldwin said. “Especially when it comes to paper-writing for the humanities, I doubt students will want to utilize AI. Yes, it’s a difficult thing to write well, but human beings will always be able to communicate better than a robot. Even if the population at large doesn’t believe that, I bet you more than anything that’s what Hillsdale students think.”

Associate Professor of English Patricia Bart said she thinks Hillsdale students will be more resistant to the new technology than most other schools.

“Nevertheless, there is a temptation to use it,” Bart said. “We need to nip this in the bud and treat it very seriously.”

The English department may need to refine its plagiarism policy, according to Bart. The current policy says work “written by another person” is considered plagiarism.

Bart said use of technology for a writing assignment should be considered “first-degree plagiarism.”

“If I found someone using ChatGPT, I would make every effort to get them expelled.”

Loading