As ChatGPT flourishes, university debates its merits

Martha Sanchez, News Editor

The new artificial intelligence chatbot Chat GPT is gaining popularity across the country and alarming some universities, who fear it could lead to more plagiarism. (Will Embree)

Students packed the class and dared the computer to compose a poem. 

“Write a limerick,” they typed. 

Easy.

“Write a sonnet,” they typed next. 

Fourteen neat lines appeared.  

“Write a shanty about Ada Lovelace,” the students said, and the cursor burst across the screen, in seconds phrasing a folk song about Lovelace, the woman who published the world’s first computer program. 

Eager, skeptical and sometimes in awe, the Tulane University students were learning the history of the digital revolution with a controversial new tool that could transform the future. 

ChatGPT, the new artificial intelligence chatbot, answers questions in clear and concise paragraphs. It responds to essay prompts. It writes poetry. And it is gaining popularity across the country, alarming some universities who fear it could lead to more plagiarism and force professors to change the way they teach.  

“Many faculty are not ready for it,” Tulane Provost Robin Forman said. “They haven’t had time to think through how this might be used as a tool in what they’re teaching.” 

For others, the chatbot means a chance to get creative and think deeper about their work. 

“Together, we’re going to learn,” Walter Isaacson, the bestselling author who teaches the digital revolution class, said, “how to use it to augment, rather than replace, our own thinking.” 

This semester, Isaacson’s students will embrace the bot. They will write papers and compare them to essays written by ChatGPT. They will critique the computer and try to beat it. Last week, Isaacson asked his students to use ChatGPT to learn about Lovelace, the 19th century programmer. So, they typed commands on their keyboards and watched the words materialize on their screens. 

Isaacson said banning ChatGPT — like some public schools did recently — would be like forcing students to stop using Google. 

“It’s very important that instead of trying to ban ChatGPT, we learn how to weave it into the way we teach,” he said. 

But passing off ChatGPT’s words as one’s own is a form of plagiarism, Rachel Champagne, director of Tulane’s academic integrity office, said.  

Cheating fears could force some professors to adapt. In-class work, drafting and revision are all areas that could see greater focus this year, Thomas Albrecht, chair of Tulane’s English department, said. Faculty could also assign spoken work, like speeches or debates. 

But a focus on process — grading outlines, revisions and drafts —  is more difficult at large universities with more students and fewer professors. 

That has led some to declare ChatGPT the end of the student essay. But Albrecht is not ready to ditch the craft. 

“I love essays,” he said. “Why would we deprive ourselves of that experience?” 

To ban or not to ban

For decades, universities have fought plagiarism and cheating. They punished ghostwriters, and they busted online paper mills. And at schools across the country, alarm bells are ringing once again. 

Will Tulane ban the chatbot? 

“No,” Forman said. Instead, Tulane will leave it up to faculty to decide the rules in each class. Some professors have sought to stop students from using it. Others are exploring its values. It is likely students will encounter professors on both sides of the debate this semester. 

“We’re all just still learning both the power and the limitations,” Forman said. “It’s not a substitute for college writing. And those who simply think it will do their work for them, I think will learn quickly that they’re mistaken.” 

The cheating question is simple. Who did the work: the student or someone else? 

ChatGPT has already fooled some. 

Academic integrity offices at schools nationwide will likely use techniques like new artificial intelligence detectors and review of older writing samples to determine if a student used ChatGPT unfairly. 

Tulane is exploring detecting software with caution because the technology is so new, Forman said. He urged students testing the tool to remain careful. 

“A lot of what it writes is nonsense,” he said. 

Experimenters and optimists

Two months ago, Jordan Kirsh-Clemenceau had no idea a computer program could do this much. 

Now, the Tulane senior is using it to create gluten- and dairy-free recipes for mac and cheese. 

It is OK, he said, to use ChatGPT to brainstorm. But he double-checks everything and analyzes the information on his own. 

“What’s really important,” Kirsh-Clemenceau said, “is making sure that what you’re creating  – the basis for it, and how you use that — it comes from you.”

Business Professor Rob Lalka is using ChatGPT in a venture class this semester too. His students use it to summarize research and riff ideas. He hopes, if ChatGPT becomes a tool of the future, their experience will teach students to use it ethically in their careers. 

Isaacson and Lalka appear to be the few professors at Tulane using Chat GPT in class. Both said they want to embrace it. 

“The whole point of technology is not replacing humans, it’s enhancing the capability of humans,” Isaacson said. “Instead of worrying about artificial intelligence, we should think about augmented intelligence, which is ways that humans and machines can work together.”

Nicholas Mattei, a computer science professor who specializes in natural language processing, said he hopes the chatbot can challenge students and professors to think more deeply about their work. 

Students will still have to write basic essays. “Sometimes you got to do things you don’t like,” Mattei said. “But it’s also on us, on the professors’ side, to try to engage more with the process.” 

That does not mean the end of the essay, he said. But with more specific prompts and greater attention to argument and detail, it could mean the end of the boring one. 

Revolution in real time

The powerful tool has pitfalls. 

“It has no logical model of the world,” Mattei said. “It will fail very basic logic puzzles.”

An example: when asked to find eight articles that discussed the pros and cons of a certain writing approach, the system gave seven references.  

None of them existed

Mattei said ChatGPT is good at basic, formulaic information that appears often on the internet. It can write eloquently about international economies. It can write songs and basic essays. And it is evolving. 

“What does that mean for our faculty? What does it mean for students?” Albrecht said. He hopes to gather the English department this semester to discuss ChatGPT. “These are the questions we’re only beginning to ask.”

And in Isaacson’s classroom, students fill seats and cram aisles to find answers. 

Leave a comment