One night in early March, Howard University senior Jamarri Kaywright was doing homework in his room when his friend said something bizarre on FaceTime.
The computer science majors were doing schoolwork when, out of frustration, Kaywright’s friend said that he was going to “ChatGPT it.”
“So I’m like, ‘what is that?’” Kaywright said, “and he was like ‘Oh ChatGPT – it’s an AI, it writes papers and stuff for you.’”
Kaywright, a D.C. native, didn’t think much of the asserted capability of this so-called ChatGPT at first, but when his friend asked it to write a story using the words “hat,” “egg,” and “boil,” the artificial intelligence (AI) technology did exactly that.
“I was shocked,” he said. “He really typed something into a machine and it just generated a response and actually gave a story. You didn’t have to edit it or anything.”
The potential that Chat Generative Pre-Trained Transformer (ChatGPT) holds is widely interrogated at Howard University and in a spectrum of different professional industries, including higher education.
ChatGPT is an AI chatbot developed by AI research laboratory OpenAI that can interact with users by having conversations, answering questions and assisting with tasks such as generating code, writing essays and summarizing texts, according to technology news website ZDNET.
The Hilltop asked ChatGPT to write a three-paragraph essay analyzing university tuitions in the United States. Here’s what it generated:
Here’s how ChatGPT works: According to Towards Data Science, the chatbot uses machine learning to create “Large Language Models” that pull information from a vast compilation of online texts, such as articles and books, to generate responses to users’ prompts.
It’s as if the AI scans text on the internet and uses the information it collects to communicate with the user.
Many Howard students and professors have been testing this technology themselves and discussing its potential impact in the classroom, including computer science Assistant Professor Dr. Noha Hazzazi.
Hazzazi believes that the potential that ChatGPT has is “really huge.” “Students,” she said, “can ask it to write papers. They can ask it even to do some programs, even write an algorithm.”
The AI chatbot has been a somewhat helpful resource for students like sophomore computer science major Angelica Stewart. Stewart has been following ChatGPT since a few days before it launched in late November, and is exploring career research opportunities in AI herself.
Although she doesn’t use it much for classes, Stewart once asked ChatGPT for help on a homework assignment. Her question was “something very complex for Computer Organization,” she said.
“I was looking everywhere on Youtube, I was using Chegg, and nothing was giving me the answer or the explanation. Then one of my classmates was like ‘You know what, let’s just try ChatGPT,’’’ Stewart said. “And then putting the question in the ChatGPT – it answered the question, like it broke it down.”
Stewart added, “I didn’t even need to go to office hours because the explanation was so profound.”
Lecturer in the English department David Stroup said there has been “plenty of chatter” about ChatGPT amongst his colleagues within the department.
Stroup teaches a variety of different writing courses and said that about 10 or 15 percent of people in each class said they have tried ChatGPT or played around with it.
Stroup pointed out that it’s “a pretty complicated question” to consider whether or not copying and pasting something generated by ChatGPT into an assignment should be considered plagiarism.
He proposed that ChatGPT is technically already plagiarizing by taking information from other sources written by people and not properly citing it. So to use a ChatGPT-generated response, Stroup believes one could argue, might as well be plagiarism.
He explained that the assignments he gives for his classes would be difficult for students to answer using ChatGPT because “I’m asking students to write about their personal experience, their personal views on how things have shaped them…,” he said. “ChatGPT can’t write about a scene from your life because it doesn’t know your life.”
After asking ChatGPT to create a writing assignment for a college composition class, it did, but “The answers sound OK,” Stroup said. “The spelling, the grammar, that’s all fine. It’s rather concise, but it speaks in generalizations. It’s not very good at inventing fake personal details.”
Nonetheless, Stroup believes that properly addressing ChatGPT “is going to be very difficult for instructors.”
The AI chatbot is already being used by students for midterm exams, for example.
Kaywright used ChatGPT himself for a Computer Organization midterm exam in mid-March.
“For the most part, it got some of the answers right,” he said. “That’s when I came to realize, ‘Oh, it can put you in the right direction for sure.’”
Kaywright said that if he had not used the AI chatbot to assist him with the exam, he “definitely” would have gotten a lower grade than he received.
“Now, when it comes to academia,” Hazazzi said, “many discussions have started to come out which are, ‘what should schools do about this?’”
Hazzazi gave her own suggestion, that “we need to have policies within the university.” She noted that it would not be acceptable to her if students were to copy and paste work from ChatGPT into an assignment or become over-reliant on the AI.
She proposed two policy options: “either lock every port with AI on it,” which she analogized to limiting Google or other websites, or say to students, “how about you use it to benefit you? How can we use it in a smarter way?”
In this regard, Hazzazi argues that a student should only use the AI in a supplemental way, for checking their work or using it to help them answer questions as examples. She stressed that “as long as they understand what ChatGPT gave out, that’s fine.”
Howard’s Provost & Chief Academic Officer, Anthony Wutoh, indicated a shared perspective with Hazzazi. “I am optimistic about the potential of ChatGPT,” he said, “but at the same time, I am concerned that it could be used inappropriately to undermine the development of critical thinking skills if not used fittingly.”
He continued by adding, “…students using inappropriately only serve to disadvantage themselves in the long run. It is a tool, and should be used in a supportive manner.”
Wutoh also said that it will be important for the university to “outline guidance regarding its appropriate use, as well as current limitations.”
At the opening of the Myrtilla Miner Building in late February, President Wayne A.I. Frederick spoke briefly about the potential of ChatGPT at Howard.
“I know that there’s a lot of fear and concern about what this would mean for our education system,” Frederick said. “I think we have to embrace these technologies, I think we have to warp and manipulate them to do what we want them to do and not become subject to them.”
ChatGPT has been subject to generating inaccurate responses to users’ questions and prompts, however, most likely because OpenAI is still in the research phase of developing the technology, according to Forbes.
For example, The Hilltop asked ChatGPT, “What happened with Howard University’s Blackburn Takeover protest?” to which the AI responded in its first sentence, “The Blackburn Takeover was a student protest that occurred at Howard University in 1968,” which is inaccurate. ChatGPT also added that the protest lasted for five days.
The Blackburn Takeover occurred in 2021 and lasted for more than 30 days, according to The Hilltop.
There seems to be a wide consensus, however, that AI like ChatGPT will strongly influence school curriculums in the future.
“I think essays, as a tool of assessment, may have to change because that’s going to be the easiest pickings for AI to kind of do the work for the student,” Stroup said.
He added, “We might see a bigger movement towards oral assessments, or some kind of interactive activity that the student has to do either in person or some way that they’re obviously not being assisted by AI.”
“It’s really growing so much, and this is an algorithm that learns,” Hazzazi said speaking of ChatGPT. “So as it learns, it’s just going to keep learning and learning, and the database will just keep getting bigger and bigger.”
OpenAI was founded in 2015 by a group of businesspeople and technology developers, including investor and current CEO Sam Altman and business magnate Elon Musk, according to MIT Technology Review.
ChatGPT is the fastest-growing consumer application in history, amassing 100 million monthly active users in January, which was months after it launched, according to Reuters. For reference, Reuters says it took TikTok about nine months to hit 100 million users, and it took Instagram two-and-a-half years.
Copy edited by Jadyn Barnett