Getting an A on a paper was easier than one University of Iowa student thought.

The UI student — who was granted anonymity by The Daily Iowan — said they had a successful history of cheating on assignments in high school. When they heard about the artificial intelligence program ChatGPT last year, they were immediately intrigued.

Chatbots are designed to answer frequently asked questions; provide information, assistance, and customer service; or even entertain the user.

The platform can explain complex topics like Albert Einstein’s theory of relativity, write an original song about the causes of the U.S. Civil War from the perspective of rap artist Lil Baby, and do everything in between — all in a matter of seconds.

The software was created by the Elon Musk-backed AI company OpenAI and was released free to the public in November 2022. By January, the platform reached 100 million users.

Like many UI students, the student who cheated with ChatGPT felt overwhelmed with the workload they had to complete for their classes. They claimed they received the prompt for a class paper a week before it was due.

“I was just scavenging through notes, lecture slides, the internet, trying to find whatever, and it was just such a hassle,” they said. “I was in a pickle at this time because I had other [school work] to worry about, and I was like, ‘I don’t have time to do this.’”

Infographic by Jami Martin-Trainor/The Daily Iowan

During this time, the student started hearing about a website that writes papers free of charge from friends around campus. The student said they had always been aware of AI services, but none as complex as ChatGPT.

“I was just like, ‘There’s no way [ChatGPT]  is possible. There’s no way this is a thing,’” they said.

The student said they were introduced to the service by a friend on campus who used ChatGPT to complete an assignment.

“I was like, ‘Well, if [ChatGPT] gave him answers for this business class, I can 100 percent plug in a random era that happened in our history,’” the student said.

The student typed the paper’s prompt into ChatGPT and was given the answer they needed for the paper.

“All I did was plug in the question, and it was perfect,” they said.

The DI tested the prompt in ChatGPT and received a 400-word paper in under three minutes. After prompting the AI further, the DI was able to get the AI-written work to be closer to 1,000 words to fit the likely parameters of the assignment and to broaden the focus of the paper.

This all took about 10 minutes.

Initially hesitant to turn in their AI-assisted paper, the student altered some of the text to be extra careful and ran the text through Google.

“I tried every single sentence, like copied it, put it in Google to see if it’ll be somehow traced back to [ChatGPT], and it’s just untraceable,” they said. “But I still had doubts because I just risked my academic promise over some stupid  [paper].”

ChatGPT pulls information from 570 gigabytes of data obtained from articles, books, Wikipedia pages, and other pieces of writing on the internet. According to ChatGPT’s homepage, the service has limited knowledge of world events after 2021.

Despite their concerns, the student turned in the paper that ChatGPT produced. They thought it had been detected, as the professor took longer than expected to grade their paper.

Finally, the student received an email informing them that their essay was graded. The grade? It was an A.

“I’m still blown away … even the ICON accuracy test [showed] it was perfect,” they said.

RELATED:  New artificial intelligence developed for radiology at the UI 

UI’s response to use of ChatGPT

Since ChatGPT was released to the masses late in the fall semester, the UI has offered a few ways to support professors.

In January, the Office of Teaching, Learning, & Technology released a guide offering instructors tips on how to talk with students about AI services.

“The newest entry in a long line of technologies that promise to disrupt higher education are artificial intelligence (AI) tools like ChatGPT,” ITS’ website stated. “Iowa students are likely to engage with these tools, and it is clear that instructors will need to consider carefully how to adapt.”

The website includes a tab that provides guidance to UI staff on how they can detect text generated by ChatGPT using platforms like GPTZero and Turnitin. Some UI professors already use these programs to review online assignment submissions.

Over 6,000 instructors from universities like Harvard and Yale have used GPTZero since its release.

Officials at Iowa State University composed a task force to study ChatGPT and other AI tools. Meanwhile, Rutgers University has left it up to professors to decide whether or not to ban AI tools or allow students to incorporate it in their assignments.

The two largest school systems in the U.S. — the Los Angeles Unified School District and the New York City Department of Education — already blocked the use of ChatGPT on school Wi-Fi because of plagiarism concerns and the chatbot’s accuracy.

After gaining national media attention following ChatGPT’s release, Charles Keene, UI Tippie College of Business associate dean for undergraduate programs and professor of instruction, said he first became familiar with the platform from his son who is a current high school senior.

Keene said his son showed him how the chatbot works and informed him that many friends his age use the platform.

Keene said it is too early for members of academia to ban ChatGPT because it would be counterproductive for future college students.

“I always think through the lens, like, if we’ve got high school students already becoming familiar with this and comfortable with it, and then they come to us, and we say, ‘You can’t use that.’ Well, they’ve been learning this,” Keene said. “It’s a part of their environment.”

Keene said ChatGPT is just the latest progression in the evolution of education technology.

“Eighty years ago, we were hand calculating everything in a statistics class, and then comes punch cards, and then along comes discs,” he said. “It’s like saying, ‘Class, we’re gonna take out our abacus today?’ well how about a calculator.”

Melissa Tully, the UI School of Journalism and Mass Communication interim director, said chatbot services pose a legitimate concern regarding students cheating.

Infographic by Jami Martin-Trainor/The Daily Iowan

“If people are using a tool like this to cheat, then they are not learning those fundamental building blocks that will be important to do a more elaborate, you know, essay or presentation,” Tully said. “Skipping the basics could be something that could be damaging if it becomes a widespread practice.”

As much as skeptics of ChatGPT want to liken it to the evil supercomputer HAL 9000 from the film “‘2001: A Space Odyssey,” the service does have its limitations.

According to the platform itself, ChatGPT may have inherent biases in its training data and also lacks:

• Emotional intelligence
• Contextual understanding of topics
• Creative problem-solving skills

Loren Glass, UI English department chair, experimented with ChatGPT and said the chatbot gives generic answers.

“I did give it a test run on some paper assignments and had some interesting discussions with colleagues about whether the resulting papers would be a B or a C, but that’s only when you give it generic questions,” Glass said.

He had it compare and contrast books and said it might get a decent grade particularly if a student turned it in for a large class and if someone tried to edit it afterward.

Glass said he immediately realized he was going to have to adapt and change some of the ways he designs class assignments.

“I used to make general paper assignments because I wanted the students to come up with their own topics, but I think now I’m going to have to do more specification to make sure that people don’t find ways to run it through [ChatGPT],” he said.

Some members of the UI English department think assignments should now be tailored to the classroom experience, Glass said. Other faculty members are trying to find ways to incorporate AI into their teaching, he said.

“Everyone acknowledges that this is a changed reality not just for teaching English, but for the culture,” he said. “But people are not in a panic about it. They’re finding ways to adapt to it.”

An AI revolution

AI chatbots like ChatGPT aren’t going anywhere. Most major technology companies are launching their own versions of the program:

• OpenAI is expected to release an updated version tool, ChatGPT 4, which is better at generating text than previous versions.
• Google recently unveiled a beta version of their rival chatbot, BardAI.
• Microsoft plans on investing $10 billion into OpenAI and announced that its search engine, Bing, is being revamped to incorporate AI to help generate user search results.
• Shares for the digital media company BuzzFeed doubled in January after the company announced plans to use ChatGPT to help produce quizzes and other content.

Keene said there is potential for AI to create new jobs for students.

“When I think about something like ChatGPT, and you ask it to write something up, and it writes it up, well, who’s going to translate that and give it context?” Keene said. “Who’s going to tell somebody how and why it said that?”

Tully said members of the UI School of Journalism will participate in a workshop conducted by the Office of Teaching, Learning, & Technology about ChatGPT and the inner workings of AI services in March.

The student who cheated said they would potentially use ChatGPT again but had doubts about how long it will remain available for students.

“I know it’s gonna get [shut] down. I hope [the UI] doesn’t respond to it because there are more things they should worry about,” they said. “I’d much rather learn at least the majority of the [material] rather than just, like, cheat my way out.”

Editor’s Note: Melissa Tully is a member of the Student Publications, Inc. Board.