November 11, 2025
SEOUL – Widespread cheating during a midterm exam at Yonsei University has reignited concerns about the lack of rules governing the use of artificial intelligence in higher education.
The incident occurred in a third-year online course with around 600 students, taught at Yonsei University in western Seoul.
The course included a midterm exam on Oct. 15 that was also conducted online. Students were instructed to record their computer screens, hands and faces throughout the exam to prevent cheating.
Despite such measures, the professor overseeing the course announced that “a large number of cheating cases were detected,” with some examinees bypassing the monitoring system by manipulating camera angles or capturing screenshots of the test and pasting them into ChatGPT — behavior later flagged as suspicious in a comprehensive video audit conducted after the exam.
According to university representatives, “around 40 students have reported their own cheating to the school.”
The controversy has drawn attention to academic institutions’ apparent lack of preparedness to address AI’s evolving role in education.
Although generative AI has been widely accessible to the public for nearly three years, university-level policies governing its use remain unorganized or close to nonexistent.
According to a 2024 survey by the Korea Research Institute for Vocational Education and Training, which surveyed 726 university students with experience of using generative AI, 91.7 percent of respondents reported using AI tools to gather information while researching or to finish assignments such as writing tasks.
“Everyone around me uses ChatGPT or Gemini to finish their university tasks or to get help in understanding course material,” a university student in Seoul told The Korea Herald on the condition of anonymity. “It’s to a point where not using the tool almost makes you feel stupid. It makes you feel like you’re lagging behind because you might spend the whole day working on something that your classmate used generative AI to finish in a few hours.”
On Everytime, an online forum used by university students in Korea, the use of AI tools to complete assignments or online exams is a common topic of discussion.
One user wrote, “I pasted my test questions into ChatGPT, and it got most of them right. It really made me feel like I wasted my time trying to solve them myself.”
Korean universities have struggled to respond.
According to the Korean Council for University Education, 77.1 percent of 131 universities lacked any formal policy on AI usage in university classrooms. Even the ones that do have guidelines give only vague instructions, such as making sure to “ensure factual accuracy when using generative AI” or to “follow instructors’ directions regarding AI use.”
Yonsei University also had similar guidelines in place regarding AI usage. However, a university official told The Korea Herald that there is “little they can do” to prevent students from “cleverly using” AI.
Instead of banning AI tools outright, experts argue that universities should develop transparent usage frameworks that incorporate AI literacy alongside students’ own critical thinking.
“A productive way to move forward from here is to allow AI use but require students to clearly cite their sources and explain their reasoning if they are answering fact-based questions in their assignments or exams using generative AI,” Kim Myuhng-joo, head of the AI Safety Institute, told The Korea Herald. “Assessments from now on need to change so that they evaluate more than just the final answer. If using AI is inevitable, universities need to adapt and test how students engage with AI.”
Kim added that there also need to be structural reforms in teaching and assessment methods that “go beyond what AI can generate,” such as making it “mandatory to give in-person presentations or participate in academic discussions.”