Illustration by Taylor Callery
As New Jersey City University professors watched, listened and conversed in a conference room in early February, they soon realized the challenges—and opportunities—that awaited them.
The group, more than 15 strong, met for a workshop on ChatGPT, the chatbot-style artificial intelligence (AI) program that’s garnered global attention since its launch in November 2022. Some in the room were already well acquainted with the technology, which was created by OpenAI, a company that specializes in AI research and development. Others were only vaguely familiar.
But when the workshop’s leader, Yufeng Wei, a chemistry and biochemistry professor and the inaugural director of NJCU’s Center for Teaching and Learning, asked ChatGPT to write an essay on Shakespeare’s Macbeth, everyone swiftly understood what the bot was capable of.
The group watched in equal parts horror and amazement as ChatGPT typed a passable essay in a matter of seconds. Those reactions only intensified as Wei demonstrated how a user—in this hypothetical case, a cheating student—can tailor results to their writing or grade level by refining the prompt entered.
“You can tell the engine who you are,” Wei explained.
“We’re doomed!” one attendee responded.
“It is becoming the Terminator!” another professor joked.
That one-liner, a reference to a movie franchise in which an AI program becomes self-aware and nukes the world, drew laughs. But it—and the overall presentation—underscored the difficulties that the pedagogical community is facing as AI becomes increasingly available.
[RELATED: Our State of Education: How NJ Schools and Students Are Faring]
Already, universities and school districts all over the world are trying to restrict students’ use of ChatGPT. The New York City Department of Education even banned the bot from its schools’ devices and networks in January. New Jersey Monthly inquired what, if anything, the Garden State might do about ChatGPT with regards to school use. While the New Jersey Department of Education declined comment for this story through a spokesperson, state educators and students willingly discussed the pros and cons of the tech and how it can be used. As many noted, ChatGPT’s potential extends far beyond cheating. “It’s a tool,” Wei said at the start of his presentation. It’s a point he made multiple times.
What is ChatGPT?
If you’re not a student or teacher, or otherwise involved in schools, you may not know what ChatGPT is. Luckily, the bot can introduce itself.
“ChatGPT is a computer program designed to hold conversations with humans,” ChatGPT wrote when we asked it for a simple explanation of what it is. “It uses advanced artificial intelligence and machine learning techniques to understand and generate human-like responses to text input. Essentially, it’s a smart AI language model that can answer questions, engage in conversation, and perform a variety of language tasks. The goal is to make the interaction with ChatGPT as similar as possible to a conversation with another person.”
In simpler terms, ChatGPT lets you enter a text prompt and gives you a text response. Prompts can range from simple tasks such as, “Tell me a joke,” to complex topics like, “Explain quantum computing.” A prompt that is more specific, such as, “Explain quantum computing as if a computer science grad student is writing,” will get a more detailed response. ChatGPT can also code, solve equations, stick to word counts, and follow formatting instructions, which is why it’s become so popular with students.
ChatGPT gets its information from the internet and other digital resources, including websites, books and articles. Its large dataset was collected and curated by OpenAI.
“If you get really good at asking the right questions, it’s gonna give you really great responses,” said Erica Holan Lucci, the assistant director of instructional design and technology services for the School of Communication and Information at Rutgers.
However, ChatGPT, which is free to use and was still in a research phase at press time, has its limitations. For starters—and pay attention, students—the information is not always factual, even if it’s written coherently. And, should your assignment require you to cite your sources, be aware that the program sometimes spits out bogus ones.
Also, sometimes it’s hard to log into ChatGPT during peak hours, when there are too many users on the site (OpenAI’s ChatGPT Plus gives users priority access, among other benefits, for $20 per month).
As of February 2023, ChatGPT had limited knowledge of world events that took place after 2021. And while the bot has been trained to decline inappropriate requests, it may occasionally provide harmful instructions or biased content, which is why the potential to disseminate misinformation is another major fear associated with AI.
For educators, however, academic integrity is the gravest concern. But, not all in the teaching profession believe ChatGPT should be banned. “Students have cheated since the dawn of time,” Holan Lucci reasoned. “If students want to cheat, they’re gonna do it. So it’s not like introducing this tool is introducing them to a new way to cheat.”
How can students & educators use ChatGPT?
Shortly after returning from Thanksgiving break last year, a Montclair High School senior noticed that use of ChatGPT had become widespread among their peers.
Some were having it write entire essays and complete other assignments, scoring passing grades without revising the bot’s output. So the senior began utilizing ChatGPT for classwork that they didn’t feel like doing, but they were unsatisfied with the results and made it a point to rephrase answers.
“I feel like it takes a lot of emotion and depth out of writing,” said the student, who requested anonymity in order to speak freely. “It’s definitely, to me at least, obvious that it is AI generated. I feel insecure and kind of guilty using it and then putting it into an assignment because it just feels like it’s not representative of my capabilities and it doesn’t sound like my voice.”
That’s not to say the student stopped leaning on ChatGPT. Rather, they changed how they use it.
The student, and nearly every educator New Jersey Monthly spoke to, said that ChatGPT can be ethically used as a powerful brainstorming tool when it comes to tasks like forming outlines and lesson plans. The anonymous student added that the bot has helped them overcome writer’s block.
ChatGPT can also be beneficial to those who speak a foreign first language and for teaching information literacy—which will soon become part of New Jersey’s K-12 curriculum—since not everything it types is accurate.
Holan Lucci, meanwhile, praised the bot’s ability to help students digest complex subjects. “I saw the potential for the tool to be used as a way to help students understand content that may be loaded with jargon that may not make sense to them,” she said.
Other professors rejoiced at the idea of ChatGPT generating time-consuming recommendation letters for them, and Wei said it could be useful for research purposes.
Still, “ChatGPT does the work for them,” argued Venessa Garcia, a criminal justice professor at NJCU.
Her colleague, biology professor Meriem Bendaoud, worried that overreliance on AI could lead to underuse of one’s own brain and poor development of critical-thinking skills and memory. She compared that to a muscle growing weaker if it’s not exercised.
But other NJCU professors pointed out that work- and thinking-reducing technology, like graphing calculators and spellcheck, have become commonplace in classrooms for decades.
“Those things are not that far away from this,” said Joseph Moskowitz, a political science professor. Moitrayee Chatterjee, who teaches computer science at NJCU, mentioned that chatbots have been around for a while now; ChatGPT is just more advanced.
AI will continue to evolve, and keeping certain tech, like smartphones and laptops, out of classrooms has proven unenforceable in the past. The way some see it, students and educators will be better off understanding and adjusting to something like ChatGPT.
“Teachers should start thinking about, how do you adapt these technologies in education rather than completely banning them like [was done in New York City public schools], which is, in my viewpoint, not sustainable,” said Princeton senior Edward Tian.
Tian continued, “It’s like, this future is going to happen. You’re kind of entering it blindly if you’re banning something rather than having a tool that sees the usage and what’s happening.”
What can be done about chatbot cheating?
Tian is offering more than just advice to educators. He’s also come up with a way to combat unethical uses of ChatGPT.
The computer science major and journalism minor—Tian is also mindful of AI’s ability to spread fake news—spent his most recent winter break developing GPTZero, a counter bot that can detect if text was written by ChatGPT. Tian, a 22-year-old from Toronto, developed his beta program at a hometown coffee shop over the course of a few days.
“The motivation for GPTZero is to build the safeguards so this technology is adopted responsibly,” Tian said, though he warns that educators and administrators should not base academic and disciplinary decisions solely on his bot.
Like ChatGPT, the most basic version of GPTZero is free and not 100 percent accurate. But Tian introduced a more precise model (GPTZeroX) for educators in late January. At the time, the product’s waiting list had signups from more than 40,000 teachers in 40 states and 30 countries. Tian also said he’s spoken to Blackboard, an education technology company, about integrating GPTZero.
Educators already use Blackboard’s SafeAssign and Turnitin services to detect non-AI plagiarism.
It wasn’t long after GPTZero’s debut that OpenAI, the creator of ChatGPT, introduced its own program for identifying AI-written text. However, the company’s announcement came with a caveat: “Our classifier is not fully reliable.”
Again, technology evolves over time.
“We don’t want ChatGPT to be used for misleading purposes in schools or anywhere else, so we’re already developing mitigations to help anyone identify text generated by that system,” an OpenAI spokesperson wrote in a statement. “We look forward to working with educators on useful solutions, and other ways to help teachers and students benefit from artificial intelligence.”
Meanwhile, there are old-fashioned methods that schools can use to curtail misuse of ChatGPT. Drastic options include limiting graded assignments to the classroom, where teachers can keep a watchful eye on students. Some have suggested combining that approach with a return to pen-and-paper assessments.
Another idea is to complicate assignments in a way that prevents or limits how much students can depend on ChatGPT. “If your assignment can be done by a bot in a few seconds, is it worth it to assign to your students?” Wei asked fellow faculty.
Others noted that knowledge is power for students and teachers alike. If students are taught about ChatGPT’s advantages and pitfalls, more may use it responsibly. “The earlier that we get into sharing with students how to properly use these tools, the better off we’ll be,” Holan Lucci said.
On the flip side, educators should make concerted efforts to familiarize themselves with their students’ original writing early in the school year or semester, if they’re not already.
“Teachers who have worked with you before or read your writing before know the quality of your work and know what it usually looks like,” the anonymous Montclair student said. “And a lot of times, because a student may copy down an auto-generated response, it sounds absolutely nothing like them. And that can be a dead giveaway.”
The student acknowledged that such use of ChatGPT “is cheating” and warrants punishment in the form of a failing grade.
Because ChatGPT is so new, many institutions have yet to craft specific policies for the AI and similar technology. “There’s a plagiarism, an honesty policy, that one could interpret as covering some of these issues,” Moskowitz said of NJCU—but is that particular enough?
Holan Lucci said that syllabi and codes of conduct should explicitly mention AI and specify the repercussions for abusing it, once such policies exist. What exactly those polices should be is still up for debate, however.
“That’s why more discussion and awareness are needed,” Bendaoud said. “Because at the end of the day, everybody’s going to approach it a different way, depending on who you are. It might be a good tool. It might be a destructive tool. It depends how we use it, but we need to be aware of it.
“It cannot be one size fits all.”
CHECK OUT CHATGPT’S WORK
Illustration by Taylor Callery
Now that you have a better understanding of ChatGPT, we want to show you just how human-like the AI’s writing is (or is not, depending on your perspective).
New Jersey Monthly asked ChatGPT to write 200 words on why students shouldn’t use ChatGPT to cheat, which you can read below. The results read like an essay, and the program does not acknowledge that it’s not always 100 percent factually accurate. But you will notice coherent sentence structures and points that are argued by real-life educators, all of which were generated in a matter of seconds.
What follows was written entirely by ChatGPT and was not edited by NJM.
As a language model trained by OpenAI, ChatGPT has been programmed to provide helpful and accurate responses to a wide range of questions. However, it is important to remember that ChatGPT is not intended to be used as a tool for cheating in academic settings.
Cheating undermines the integrity of the educational system and can have serious consequences for students who are caught. In addition to the ethical concerns, cheating can also have long-term negative effects on a student’s academic and professional career. By using ChatGPT to cheat, students are not only cheating themselves out of an opportunity to learn and grow, but they are also risking their academic and professional future.
Furthermore, using ChatGPT to cheat can also be detrimental to a student’s personal growth and development. The process of learning and problem-solving is an essential part of a student’s educational experience, and by relying on ChatGPT to provide answers, students are depriving themselves of the opportunity to develop critical thinking skills and problem-solving abilities.
Ultimately, the decision to use ChatGPT to cheat is a short-sighted and self-defeating strategy that can have serious consequences for a student’s academic and professional future. Rather than relying on ChatGPT to provide answers, students should focus on developing their own skills and knowledge through hard work, perseverance and dedication.
No one knows New Jersey like we do. Sign up for one of our free newsletters here. Want a print magazine mailed to you? Purchase an issue from our online store.