South Korea's prestigious "SKY" universities – Seoul National, Korea, and Yonsei – are currently facing a significant challenge: a surge in AI-assisted cheating cases. While institutions are scrambling to respond, their efforts often remain fragmented, focusing on individual incidents rather than establishing comprehensive, campus-wide policies for AI use and online learning. This piecemeal approach highlights a broader struggle within academia to adapt to the rapid advancements of artificial intelligence.
Korea University's Struggle with Exam Integrity
Korea University, for instance, detected AI-assisted misconduct in a general education course's midterm exam, leading to its voidance. For the upcoming final, the university is implementing stringent measures. The number of questions will drastically increase from 35 to approximately 100, each accompanied by strict time limits. Furthermore, a randomized question order system is being considered to prevent large-scale online answer sharing. To bolster security, "Trust Lock" software is planned for activation, blocking screen sharing, recording, and unauthorized programs, complemented by random live monitoring during exams.
However, these robust measures are currently limited to the problematic class. The university has yet to decide on campus-wide adoption of tools like Trust Lock, citing difficulties in disclosing internal decision-making processes regarding AI and online learning policies. An investigation into the students involved is ongoing, but identifying all participants in a course with over 1,400 students spread across numerous anonymous online chat rooms presents a formidable challenge, raising doubts about the likelihood of widespread disciplinary action.
Adding to the urgency, another incident of collective cheating emerged at Korea University during an online engineering quiz. Students exploited a system loophole, allowing them to retake the quiz after checking answers, with some reportedly using generative AI tools like ChatGPT.
Yonsei and Seoul National's Responses
Yonsei University, which also uncovered an AI-related cheating case, has had AI use guidelines in place since last year, updated in September. These guidelines provide basic principles, such as fact-checking when using generative AI, and instruct students to follow course-specific policies. Yet, university officials admit these are more recommendations than enforceable rules, indicating a gap in institutional enforcement.
Seoul National University, upon detecting cheating in a class section, opted for a retest, concluding it was individual misconduct rather than large-scale coordination. The university has been proactive in developing AI ethics guidelines since forming a task force in August, engaging faculty, students, and external experts in the process.
The Urgent Need for Unified AI Policy
The ongoing incidents underscore a critical need for universities to forge a unified strategy. Song Ki-chang, an honorary professor of education at Sookmyung Women’s University, emphasized this point, stating, “I think it’s time for universities to reach a broad agreement on how AI should be used and what its limits should be.” Without clear, enforceable, and campus-wide policies, educational institutions risk being outpaced by technological advancements, compromising academic integrity and the fairness of assessment.