At UChicago and Beyond, ChatGPT Sparks Debates about Cheating in College Classrooms

Following ChatGPT’s release in November 2022, professors at UChicago and across the nation are searching for ways to prevent academic dishonesty in their classrooms.


By Naina Purushothaman, Senior News Reporter

The popularity of ChatGPT, an artificial intelligence (AI) chatbot known for its detailed responses to human prompts, has resulted in heated discussions on many college campuses surrounding the potential for its misuse—namely, to cheat.

ChatGPT was launched in November 2022 by the research laboratory OpenAI. According to OpenAI, ChatGPT is a chatbot that interacts in a conversational way. It uses a large language model that has been fine-tuned over time. The technology’s versatility and ability to source information from a large variety of sources have led it to rapidly become popular. In January, the site reached the 100 million user milestone. However, the rise of ChatGPT has some universities concerned about the possibility of using the tech for cheating. 

For example, faculty at Tufts University are split on whether ChatGPT should be used in the classroom. According to The Tufts Daily, some professors view usage of ChatGPT as academic dishonesty, while others see ChatGPT as an opportunity to revise syllabi in order to engage students in novel ways. 

At UChicago, political science professor Ruth Bloch Rubin, computer science (CS) professor Borja Sotomayor, and Executive Director of the Chicago Center for Teaching and Learning Robin Paige spoke to The Maroon about their thoughts on the technology and unique ways instructors can reduce its usage for cheating.

Bloch Rubin has already begun considering how ChatGPT could affect her classes, such as “Introduction to American Politics,” which she taught in the winter quarter. Bloch Rubin noticed that ChatGPT is good at aggregating content to produce coherent arguments. 

“I wouldn’t say the concern is that ChatGPT can mimic exceptional or very good student work product,” she said. “I think the challenge is distinguishing between ChatGPT and more mediocre work.”

Bloch Rubin has heard about some ways in which professors can counter ChatGPT. However, those strategies come with trade-offs. 

“I think the general advice has been to avoid take-home exams and essays, but I think that does students a disservice,” Bloch Rubin said. “One of the most important things a social science curriculum can teach students is how to communicate ideas about data or arguments made by others in writing. That’s a skill that is marketable in nearly every profession. So limiting essays and prioritizing in-class writing or evaluative assignments seems counterproductive, not to mention anti-intellectual.”

Bloch Rubin has adopted a strategy to counter ChatGPT instead of cutting assignments from her syllabi. 

“I’ve accepted that some students may experiment with ChatGPT. But I want to avoid rewarding those who turn in a ChatGPT product as their own,” she told The Maroon in an email. “So I’ve tried to craft my assignments to confuse the AI or generate very bad output. In practice, this means creating elaborate essay prompts with fictional details specific to the prompt and asking for recommendations/assessments based on those facts.” 

ChatGPT has also raised several questions within the computer science department. Sotomayor believes ChatGPT can undermine introductory CS courses, but he does not believe the more advanced courses are at risk. 

“While ChatGPT may be an expedient way to solve a problem and get credit for it, you’re missing out on the really valuable part of working on the problem: practicing your programming skills,” Sotomayor said. “Using ChatGPT in this way in an introductory course is like trying to learn how to play the piano by having a robot sit next to you and play the piano for you.” 

Sotomayor also said the use of ChatGPT should not be banned entirely. “I don’t think we should try to ban ChatGPT, particularly in CS, where students will be allowed (and even encouraged) to use such systems in a professional setting,” he said. As for using AI for class assignments, he said that while students cannot use it to implement entire or substantial parts of a project, “consulting” AI is allowed. 

For example, the policy given to a final project in CMSC 14200 reads: “You must not use AI systems, such as ChatGPT or GitHub Copilot, to write or generate any of your code. Using those systems in a manner similar to consulting online resources (e.g., asking ChatGPT something like ‘How do I play a sound using pygame?’) is technically acceptable, but you should take any answer with a healthy dose of salt, as these systems will sometimes provide unequivocally wrong answers.” 

Instead of focusing on academic policies relating to ChatGPT, Sotomayor believes the issue at the root of academic dishonesty is a much broader problem. “I think the focus should be less on coming up with ChatGPT-specific policies, and more on addressing a larger issue: the excessive focus on grades, which leads students to focus less on their learning and more on finding the path of least resistance to an A,” he said.

To remedy this problem, the CS department has begun experimenting with “specifications grading,” which focuses on giving students substantive feedback instead of percentage grades and allows them to improve on their mistakes. 

“These kinds of grading schemes tend to allow students to focus more on their learning, reducing the incentives for cheating. Our experience with it has been, so far, pretty positive,” Sotomayor said. 

Paige agrees with Sotomayor that the temptation for students to engage in academic dishonesty comes from a larger problem than the introduction of ChatGPT. “We must recognize that most students consider violating academic integrity policies (using ChatGPT or another means) when they are experiencing an unhealthy amount of stress, they are unclear about expectations, and/or they don’t see the value in what they are learning,” she told The Maroon in an email. 

To discourage the use of ChatGPT, Paige believes that professors should continue their dedication to creating an authentic learning environment, a piece of advice she would give not just for ChatGPT prevention but also for maximizing student engagement as a whole. She said actions such as providing prompts that are transparent about the learning goals, scaffolding major assignments, and providing students with feedback at several stages throughout the assignment process could discourage academic dishonesty overall.

“These practices encourage authentic learning in all circumstances,” she said.