College students ready to embrace ChatGPT
Many college students are approaching artificial intelligence with optimism and a willingness to embrace how the technology can enhance their learning, despite the consternation it has caused educators.
ChatGPT, a free AI technology made public in November, has pushed a debate on using AI in education to the mainstream. The platform works as a chatbot, giving human-like responses to questions and tasks it is asked to complete. Students might then use the responses to supplement their work — or even complete their work in full.
For educators, this has sparked concerns of cheating and a loss of critical learning skills their students need. In some K-12 public schools, ChatGPT has been blocked. In some colleges, professors have changed how assignments work to limit opportunities ChatGPT could be used for cheating.
College students are at the beginning of discovering the capabilities of ChatGPT and many are optimistic, but still cautious about what the future could hold.
College students still learning how to use ChatGPT
Despite the alarms ringing among educators, ChatGPT is at the beginning stages of its popularity, and not all students understand the full extent of its capabilities.
Will Downey, a 20-year-old student from Virginia Tech in the computer science department, said many of his peers know of ChatGPT and have used it, but it isn’t as widespread on the rest of campus.
“I have a different perspective than most people since I’m in the computer science department. We tend to have a faster pickup of like, different technologies and newer technology. So you kind of see those spread through a little faster,” he told The Hill.
He said he expects it to pick up more traction as students talk and get more into coursework this semester.
After discovering ChatGPT through word of mouth or social media, many people dabble with the technology by asking it to complete silly tasks such as writing a song.
“I remember I was using it at home just messing around with it, seeing what it could do. I was hearing all these crazy things about how incredibly realistic and how quick it was just pumping out flawless responses, so I messed with it a little bit. I haven’t used it a ton for any sort of practical matters, but I’ve played around with different things that I can do,” said Dace Potas, a senior political science student at DePaul University in Chicago.
Embracing ChatGPT for school
Andrew Rockovich, a 23-year-old physics graduate student at Ohio State University, said he was introduced to ChatGPT by a friend a few days after its creation. The technology was immediately able to help him with trouble he was finding in his coursework.
Rockovich was struggling to write code for an experiment he was working on because it “seemed like a thousand web pages full of documentation on how to use a device I knew nothing about.”
While it could have taken him two days to read the documentation he needed, he was able to work with it on ChatGPT and “even able to tell it to use a specific code libraries and what types of data I wanted to collect.”
Although ChatGPT did give him “a lot of small errors,” he said the “big picture stuff got me pointed in the right direction, and I was able to work out the bugs in a small amount of time.”
Along with pointing students in the right direction, ChatGPT could help students get basic questions answered.
“I’m a little bit excited because sometimes when I have a question, and it’s a simple question, I could just ask it the chat GPT and get a quality answer, rather than having to go in person to office hours or wait on a TA and bother them,” Downey said.
Do benefits outweigh the risks?
College students do acknowledge the risks and concerns educators have raised about ChatGPT, saying it would be hard to not imagine some students using the technology in unethical ways.
“Kids are always going to find a way to cheat,” Potas said, adding that he believes new technologies will be created to detect if someone turns in an assignment partially or fully written by AI.
OpenAI, the company that created ChatGPT, previously said in a public statement it was working with schools to help address concerns about cheating.
But some college students seem to think the benefits given to learning through AI outweigh the risks.
“I do think it does have a lot of like really good uses, and I do think professors should embrace it,” Virginia Tech’s Downey said.
He said he has already seen professors coming up with “unique tactics” to eliminate cheating, such as moving assignments to be in-person or finding programs that can detect when bots such as ChatGPT are used.
While Rockovich says he hasn’t seen teachers at his school really address ChatGPT, he also said he thinks it is in the best interest of professors to work with it.
“I think it is all around a very beneficial tool for my field, and people in general. If anything, you constantly are thinking you have to double-check answers it gives, so it makes you think more critically than just searching and finding a solution somebody else worked out online,” he said.
Concerns for the future
Some students realize the technology does have pitfalls that need to be addressed.
Jason Chiulli, an 18-year-old sophomore at American University, said there are real concerns about academic integrity, especially as technology like ChatGPT gets smarter.
ChatGPT is a learning language model, meaning that the more people use the program, the smarter it will get.
“With millions of people using these websites, the AI becomes smarter and it becomes better. So in the future, ChatGPT might be able to output much longer paragraphs, much longer stories, much longer essays, which will kind of further put academics into danger,” Chiulli said.
Technologies are in the works to be able to detect when students are cheating with ChatGPT, but teachers may still adjust assignments and how they will conduct classes to mitigate cheating.
Chiulli said the concept of having to move all essays in person is a “concern” of his for the future. He said he is also unsure that as the AI technology gets better that there will be sufficient technology to detect when it is used.
“I don’t know if that kind of software will be able to detect more advanced AI writing in the near future,” he said.