ChatGPT and its ability to generate original text from almost any user-provided prompt has cast some uncertainties about the future of technology use in and out of the classroom at Gunn.

The artificial intelligence chatbot developed by the startup OpenAI was originally launched as a prototype in November 2022. Since its official release in February, its popularity has skyrocketed. ChatGPT’s ability to output detailed responses to almost any question has made it appealing to some students as a completely automated homework-completing machine.

The Gunn student handbook defines what does and does not count as academic dishonesty: “Allowing others to complete your course work or to take your quiz, test and exams is considered cheating and could result in a review by your teacher followed by consequences.” English teacher Diane Ichikawa believes that using ChatGPT to complete school assignments violates that honor code. “I showed my freshmen a story where a man bought a painting from an artist, and then interviewed the artist saying that ‘I bought this under the idea that you had painted it, but it turns out it was AI.’ And the artist said, ‘Well, but I put the prompt in for the AI,’” she said. “All the students laughed at it, and I said, ‘If you put the prompt in for an essay, and it spits out an essay, did you write it?’ I think it’s clear to students that it’s not actually their own work—we all know that it’s a shortcut.”

Computer science teacher Joshua Paley compared students using ChatGPT to do schoolwork with the online school experience of the 2020-2021 school year. “During the pandemic, students didn’t have ChatGPT, but they did have Discord,” he said. “So imagine you’re a teacher and you’re giving a test during the pandemic. How long do you figure it will take for the test to be visible by all students on Discord and for them to be chatting about it?”

English teacher Justin Brown is still considering what to do about the rise of ChatGPT usage among students. “Right now we’re just in the stage of trying to get to know the technology and what it can and can’t do, as well as worrying about how much we should really change what we currently do,” he said. “The important thing we’re trying to figure out is how we can use (ChatGPT) to enhance what we do rather than have it be an obstacle that’s a problem for us.”

Ichikawa has done several activities with her students involving ChatGPT, with the goal of demonstrating the gap between human- and machine-generated work. “I had students write about ‘The Grapes of Wrath’ for about 20 minutes,” she said. “Then I had them plug the same prompt into ChatGPT. We compared the two responses and saw quality and depth differences.” Ichikawa did another activity with her freshmen students to highlight faults in the AI. “My freshmen plugged in some poetry prompts, and even with very specific prompting to not rhyme or meter the lines, it still continued to rhyme and meter everything, so they could see some of its limitations,” she said.

Although faults do exist with the software, junior Om Mahesh believes that ChatGPT can be a useful tool for students with specific questions. “If you’re trying to search for something like a synonym for some word, ChatGPT is pretty good at that,” he said. “I think it’s just a better way of Googling.”

That being said, Mahesh understands that ChatGPT is far from all-knowing and struggles to be helpful in many cases regarding schoolwork. “There was one time when I tried to ask a math question like, ‘What is the prime factorization of one?’ and it gave me a wrong answer,” he said. “ChatGPT is just really bad at math. If you try to do five-digit addition with it, it just can’t.”

Ichikawa agrees that the technology can be helpful for students when used with caution. “The ability to ask (ChatGPT) for prompts, like ‘What should I write about?’ or (when) it’s used as a brainstorming generative tool could be helpful,” she said.

Others, such as freshman Meilin Hansen, believe that using ChatGPT for schoolwork causes more harm than good. “If you’re relying on it to do your homework, then that’s a very strong dependence on a computer that you can teach two plus two is five,” she said.

Still, she expects many students to attempt to be academically dishonest on assignments by using ChatGPT to complete them. “I anticipate students using this to cheat on all sorts of things,” Hansen said. “I’ve had to resist strong temptation to not.” Mahesh echoed her thoughts. “Cheating on essays is definitely spreading and becoming more and more common,” he said.

Despite the potential for academic dishonesty, Paley doesn’t plan on adjusting his classroom routine to accommodate for the existence of ChatGPT. “At the end of the day, if you use ChatGPT to do the programs in my class, fine,” he said. “Have fun on the tests that are on paper.”

Paley also believes that, like him, other teachers won’t make any drastic changes to their classes to prevent academic dishonesty stemming from ChatGPT. “I don’t think that there are many teachers out there who are interested in policing ChatGPT,” he said. “Nobody wants to deal with that, and I know I don’t want to deal with that.”

ChatGPT’s computational power, precision and accuracy can only grow over time as its developers continue to improve it. Ichikawa accepts that academic dishonesty from ChatGPT and similar applications will be an ongoing issue for many years to come. “It’s only going to get better and more sophisticated, so I don’t think that we should bang our heads against the wall trying to stop it,” she said. “There should be ways that we can try to work with it. I don’t know what they are just yet.”

Regardless of whether or not overarching Gunn policies or individual classroom policies change as a result of ChatGPT, its existence will most likely be permanent. While Paley doesn’t see himself enforcing policies regarding AI in his classes at all, he still advises awareness of the issue. “The important thing to understand is that ChatGPT isn’t going to go away,” he said. “That’s the world we’re all stuck in.”