‘AI cheating’ shook the university… A gray area where no one is responsible [AI attack]

Universities, students, and professors are all confused by AI cheating... Responsibility structure unclear
Expert: “It’s not a ban, it’s a standard… We need to set a boundary between learning using AI and cheating.”

Since the spread of generative AI, cheating at universities is rapidly increasing, but the system and responsibility structure to regulate it are still vague. It is pointed out that the ‘gray area’ is growing in a structure where responsibility is not clearly attributed to universities, students, or professors.

AI cheating in college… 224 cases detected over 5 years

Recently, in a non-face-to-face lecture at Yonsei University, some students committed group cheating using AI during the midterm exam. In an online exam with about 600 students, test takers had to submit a video showing the screen, hands, and face, but some students evaded control by manipulating the shooting angle or running multiple programs on the screen. At Seoul National University, cheating using ChatGPT was discovered and a re-exam was conducted, and at Korea University, group cheating was confirmed through a KakaoTalk open chat room.

According to data received from the Ministry of Education by Kang Kyeong-sook, a member of the Korea Innovation Party, there were 224 cases of cheating detected at 49 universities across the country over the past five years (2020 to November 17, 2025). There were 65 cases of sharing problems and answers, 48 ​​cases of using unauthorized data, 41 cases of traditional cheating, and 40 cases of using electronic devices, and there were 4 cases of ‘Chat GPT specified’. This includes only cases directly recognized by the university headquarters, so the actual scale is likely to be larger.

Inadequate university response… Guideline limits

Universities apply disciplinary measures such as F grades, ethics education, and community service to students who are caught, but the response system is inadequate. This is because although the use of AI has increased explosively, revision of school regulations and establishment of AI use standards have been slow.

In a survey by the Korea Research Institute for Vocational Education and Training, 91.7% of college students (4-6 year universities) last year answered that they use AI for assignments and searches, but in a survey by the Korea Council for University Education, 71.1% of 131 universities nationwide did not even have AI-related guidelines. Even universities that have guidelines are assessed to have limitations in actual application as they are limited to declarative statements or comprehensive expressions.

Universities are aware of these problems, but their stance is that they have no choice but to remain at the level of guideline guidance for now. Yonsei University said, “The authority to evaluate classes lies with professors, so it is difficult to force or determine whether or not to use AI at the headquarters level,” and added, “We only provide evaluation standards and guidelines for AI use.” Seoul National University formed the ‘AI Ethics Guidelines Establishment TF’ in August of this year and is discussing establishing standards.

“Test/assignment verification system” needs to be redesigned

Some believe that this confusion is also related to the structure and operation of the university. Although non-face-to-face lectures have increased rapidly since COVID-19, criticism is that online evaluation systems such as test management and ethics education have not been sufficiently established, increasing the possibility of AI abuse.

Experts emphasize that this incident should be used as a starting point for redesigning the evaluation system for the AI ​​era. Park Nam-ki, professor emeritus at Gwangju National University of Education, said, “The boundary between legitimate AI use and misconduct must be clear,” and added, “For AI-use tasks, an evaluation structure and procedure that proves the process, such as a written plan, step-by-step review, and supporting materials, is needed.”
 
Legal industry: “AI use standards and procedures must be stipulated”

The legal community emphasized the urgency of legal and institutional response, saying that the standards of responsibility and level of punishment are different for each university, so sanctions for the same behavior are mixed.

Jae-gyu Yoo, a lawyer from the Pacific AI team at Law Firm, said, “In a situation where AI has become routine, a response that focuses on banning it is not effective,” and added, “The scope of AI use allowed in testing must be clearly defined and provided with sufficient guidance in advance.” Attorney Kang Jeong-hee said, “We need to establish a system to evaluate the achievements in using AI.”

Jeong-won Heo, lawyer at Daeryun Law Firm, said, “The key is whether the university clearly notified in advance whether or not AI will be used and whether the student was intentional.” He added, “Disciplinary standards and procedures should be stipulated in the school regulations, and CBT-type testing equipment or internal network-based evaluation systems should also be reviewed.”

In a situation where it is not even clear whether AI is a tool or a means of cheating in the educational field, the university community is at a crossroads where new standards must be established.

Reporter Hanna Kim hanna7@kukinews.com

[View full article]
‘AI cheating’ shook universities… A gray area where no one is responsible [AI attack] (Shortcut)