SALT LAKE CITY — Artificial intelligence is now in the classroom helping teachers teach and students learn, but AI also has a dark side. 

Since the interactive chatbot ChatGPT — an artificial intelligence tool from the company OpenAI — launched in November, teachers everywhere have been distressed about how it can facilitate cheating by students.

After all, the chatbot can write and debug computer programs, compose music, teleplays, fairy tale and student essays; answer test questions (sometimes, depending on the test, at a level above the average human test-taker) and write poetry and song lyrics, according to Wikipedia.

According to OpenAI guest researcher Scott Aaronson, OpenAI is working on a tool to attempt to watermark its text generation systems so as to combat bad actors using the chatbot for academic plagiarism or spam.

‘Everybody is cheating’: Why this teacher has adopted an open ChatGPT policy

Brainstorming with AI

Utah Valley University’s Armen Ilikchyan, an associate professor of technology assessment and innovation management, joins KSL NewsRadio’s Dave Noriega and Debbie Dujanovic to talk about AI on campus.

“I’m 100 percent sure our students are experimenting with [ChatGPT] on a daily basis. And yeah, so it’s around, and we know about it, and we are getting prepared for this new age of education with AI,” Ilikchyan said.

“OK, so obviously my mind goes to the cheating element of it. Is there a positive use for it? Is there a way that students can use ChatGPT in a way that actually enhances their experience in college?” Dave asked.

Ilikchyan said the AI tool is good at generating ideas or brainstorming.

“Now it doesn’t think and doesn’t care about ideas,” he replied. “It can spit out hundreds of ideas by just pretty much random generation. And then it’s up to us to go through that and think, Oh, that might be interesting. Why don’t we explore this idea?”

Detecting what a student has used ChatGPT for

“Is there a way to detect if a student has used ChatGPT to create an entire essay or paper that they’ve turned in? Or are kids getting away with it at this point?” Debbie asked.

Ilikchyan mentioned OpenAI’s planned watermark feature (referenced above).

“Every text that will be generated by AI will be marked and so can be easily detected. However, I think the focus shouldn’t be on detecting. We have to completely rethink how we do our assignments,” he said. 

The associate professor said the best way forward for students and teachers is using ChatGPT to aid in education.

“We have to go to the basics and think about why are we in that classroom? What are we trying to teach our students and then go in and redo the assignments so that there is no need to —  we don’t want to fight technology. We want to incorporate that. We want to embrace technology and move on and get the benefits of using technology,” Ilikchyan said.

 

Dave & Dujanovic can be heard weekdays from 9 a.m. to noon. on KSL NewsRadio. Users can find the show on the KSL NewsRadio website and app, as well as Apple Podcasts and Google Play.