In order to prevent students from using ChatGPT to cheat, some university professors consider returning to pencil-and-paper exams and handwritten essays
CTOnews.com The development of artificial intelligence (AI) brings new challenges to education. Fortune magazine recently reported that some university professors decided to use paper-and-pencil tests and handwritten essays to evaluate students in the fall semester in order to prevent them from cheating with OpenAI's ChatGPT.
ChatGPT is an AI program capable of generating natural language text, developed by OpenAI. The program has demonstrated amazing ability in multiple areas, such as getting top marks on AP biology tests and completing freshman year courses at Harvard with a 3.34 GPA.
CTOnews.com notes that since its launch, teachers, administrators and students have questioned the role of AI in education. Some schools choose to ban ChatGPT outright, while others explore how to make it a learning tool. As summer draws to a close, however, some college professors are looking for ways to counter generative AI and make their exams "immune to ChatGPT."
"It's no longer possible to ask students questions like 'Explain in three sentences what the Krebs cycle in chemistry is,' because ChatGPT gives a perfect answer," Bill Hart-Davidson, associate dean of Michigan State University's College of Arts and Sciences, told Fortune.
Some worried professors say they plan to return to handwritten assignments and oral exams to avoid students using generative AI. "I intend to take a medieval approach to students and go back to oral testing," Christopher Bartel, a philosophy professor at Appalachian State University, said in January. "They can take notes with AI-generated text, but if they can say it, that's another story. "
A Canadian writing professor says he plans to personalize assignments to reduce ChatGPT use.
ChatGPT usage, meanwhile, dropped nearly 10 percent from May to June, which some techies believe is because most students are on summer break. If students are the main users of the program, experts believe this could mean OpenAI has limited value, a narrow user base, and few usage scenarios.
