ChatGPT is already here – what is NYU doing about it?

“The existence of a well-written paragraph is no longer evidence of human effort,” reads a memo to NYU schools warning them of the plagiarism opportunities that ChatGPT and AI programs like it bring to the classroom.

ChatGPT, a readily available, superhuman artificial intelligence program that can quickly generate text from a prompt, has caused a stir at universities across the country. Faculty and administrators disagree as to whether it can be used as a teaching tool or if it is a plagiarism tool that has no place on a college campus. NYU doesn’t yet have a universal policy that covers the use of AI tools in the classroom, but some individual professors have included recommendations in their curriculum for the first time this semester.

As part of its response to AI, NYU has also created a task force of 50 faculty and staff to help faculty get guidance on ChatGPT. Clay Shirky, vice president of educational technology who leads the group, said the decision to create it came after some educators discovered their students were using AI as early as December last year.

“The technology here isn’t about what’s new – it’s revolutionary, essentially, about usability,” Shirky said. “Almost everyone understands that this change has already happened. Teachers are now adapting, and this adaptation will, in the long run, involve talking to students class by class.”

This semester will serve as an experiment on using ChatGPT in the classroom at NYU, Shirky said, and the results will be used with students and faculty to determine recommendations for next semester. He added that the rules governing the use of AI in the classroom will change and evolve over the next few years, reflecting the breakneck pace of technology development.

On January 21st, the Chancellor’s office, which is responsible for administering academic standards at the university, sent a preliminary memo outlining recommendations for adapting classes to ChatGPT’s existence. The guidelines, which focus on writing assignments, give educators three possible responses: looking for the use of AI in completed classwork, creating assignments that are difficult to complete with AI, or allowing the use of AI for restricted coursework.

“For writing educators, the essential change is that the existence of a well-written paragraph is no longer evidence of human effort,” the note says. “There is no short-term solution to this revolution in the usability of these tools. Instead, we will gradually but continually adapt our practices over time.”

Other universities use a similar approach, preferring to adapt to AI tools rather than forbid their use by students. One professor at the Wharton School of the University of Pennsylvania even requires students use ChatGPT, although it requires them to disclose where and how the AI ​​was used.

The note also advised educators to exercise caution when using programs designed to detect text written by ChatGPT, explaining that they are sometimes inaccurate. OpenAI, the company behind ChatGPT, described proprietary AI discovery tool as imperfect.

In one of the French courses at the College of Arts and Sciences, all essays must be completed during class to prevent the use of ChatGPT.

Reva Sharma, a freshman on the course, said the change made it harder to work on the course. However, her other professors are more open-minded about the use of AI in the classroom—in one of her writing classes, a professor asked students to compare their writing to what AI uses.

“My teachers said we shouldn’t use it to write our real essays, but today we used it in class,” Sharma said. “We all had to choose a movie to write an essay. She first asked us to write a two-sentence movie synopsis individually, and then asked ChatGPT to do the same for all of our films.”

According to Richard Xu, teaching assistant for the CAS course in computer science, the professors began adapting their courses to ChatGPT even before this semester. Xu said the class instructor launched the fall final exam questions via ChatGPT to ensure the program couldn’t solve them.

Dino Saussi, a professor who teaches the AI ​​course at the School of Professional Studies, said it allows students to use AI tools in addition to their classroom work. At the same time, he said he was making changes to his assignment hints, requiring them to be more personal – demanding details that the AI ​​wouldn’t know. He hopes that in this way students will use both technology and their own experience in their work.

“NYU students are creative, hard-working and ambitious,” Saussy said. “If we can help them think of ChatGPT not only as a valuable tool for learning about academic disciplines, but as a vehicle for learning about the world and the ethical implications of new technologies, we’ve done our job.”

While NYU’s Academic Integrity Policy does not explicitly address the use of ChatGPT in coursework, it does prohibit students from violating school and department-specific academic policies. It also prohibits students from submitting work they did not create as their own.

Robert Ausch, a professor teaching the Cognition course in the psychology department this semester, said his department has yet to release a guide to using ChatGPT in the classroom, but the program hasn’t changed the way he teaches his class. Aush said that he thinks it will be difficult for a student to use ChatGPT in his course because his exams are very specific to what he discusses in lectures.

“I don’t think it’s the job of teachers to ‘trap’ students — I don’t think of my students as kids to look after,” Aush said. “Life will eventually teach them that some shortcuts can sometimes work, but in the end, if you can’t do what you’re supposed to do, the world will notice.”

Contact Carmo Moniz at [email protected]