This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Multiple elite universities in South Korea (the 'SKY' universities) reported mass cheating incidents during October 2025 midterm exams. Students at Yonsei, Korea University, and Seoul National University were found to have used ChatGPT and other AI tools to complete online tests in violation of explicit bans. The incidents resulted in disciplinary actions, including failing grades and the requirement for exam retakes, highlighting ongoing challenges in academic integrity within higher education.
Reports from South Korea's SKY universities indicate multiple AI-assisted cheating incidents during the October 2025 midterms. At Yonsei, dozens reportedly used ChatGPT and other tools despite bans in an Oct. 15 online exam. Korea University and Seoul National reportedly found students using AI or sharing answers via group chats, prompting punitive sanctions across the institutions.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Students used ChatGPT to cheat on university exams to gain an unfair academic advantage, which directly aligns with AI-facilitated plagiarism and cheating under sub-domain 4.3.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by students intentionally choosing to use ChatGPT to cheat on exams after the AI system was deployed.
Students at elite South Korean universities used ChatGPT to cheat on midterm exams. While highlighting ongoing challenges in academic integrity and educational policy, the incident carries negligible national security implications.
The report describes students using ChatGPT as an assistant to help them write answers, translate text, and find research papers, which fits the Assistant profile of augmenting human capabilities with human oversight.
Limited Risk: The system involved is ChatGPT, which is a chatbot and general-purpose AI, falling under Level 3 requiring transparency obligations.
ChatGPT was used by students as a writing assistant and question-answering tool to generate exam answers and translate text.
The report clearly details the events, the universities involved, the AI tool used (ChatGPT), and the consequences (failing grades, exam retakes). There is little ambiguity about what transpired.
None
Technology
Widespread ChatGPT Availability
OpenAI's ChatGPT release made advanced AI tools easily accessible to students.
AI-Powered Answer Sharing Apps
Students used phone apps with AI features to share test answers.
Human Factors
Hypercompetitive Academic Culture
Intense pressure to succeed at elite SKY universities drove students to cheat.
Breach of Trust and Ethics
Students violated explicit agreements with professors not to use AI tools.
Peer Collaboration to Cheat
Students active in group chats shared screens and answers during exams.
Process and Methods
Vulnerable Online Exam Format
Administering exams online created opportunities for unmonitored AI usage.
Outdated Assessment Methods
Testing focused on information recall which is easily replicated by AI.
Inadequate Proctoring Methods
Laptop cameras and basic proctoring failed to detect real-time AI usage.
Management
Inconsistent AI Policies
Professors held conflicting views on whether to allow or ban AI tools.
Slow Institutional Adaptation
Colleges failed to update curriculum and testing to keep pace with AI.
Regulatory Environment
Unenforceable AI Guidelines
University guidelines exist but lack effective mechanisms for enforcement.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
