AJP Asia Intelligence Agency

Education Minister Resigns Amid AI Cheating Scandal, Calls for College Entrance Exam Reform Rise


2027 College Scholastic Ability Test (CSAT) has seen numerous instances of cheating using artificial intelligence (AI) and ultra-small smart devices, leading to the abrupt resignation of Education Minister Choi Kyo-jin on November 22. This has reignited discussions on reforming the CSAT. President Lee Jae-myung has called for a comprehensive overhaul of the national examination system, and the National Education Commission is reportedly considering expanding written and essay evaluations. This raises questions about whether the multiple-choice format, which has been in place for over 30 years, will undergo significant changes.

In a statement released through the Education Ministry's spokesperson, Minister Choi expressed, "With the exam approaching, various changes have caused inconvenience to examinees and the public," adding, "I cannot escape criticism for not being prepared in advance." President Lee accepted Choi's resignation about an hour after it was submitted.

On November 20, President Lee had ordered a strict response to threats against the fairness of the exam during a meeting with senior aides and via social media, emphasizing the need for a complete reform of the national examination system. The early resignation of the head of the relevant ministry, coupled with the president's commitment to reform, has led to expectations that discussions on changing college admissions evaluation methods will gain momentum.
Ultra-small AI devices breach multiple-choice format, reigniting discussions on written and essay evaluations
Recent events have starkly revealed that advanced wearable devices, such as smart glasses and ultra-small cameras and earphones combined with generative AI, can undermine the analog monitoring systems in examination halls. The current multiple-choice structure, which requires selecting one answer, is structurally vulnerable to cheating facilitated by advanced devices, as even a short number or symbol sent from outside can easily lead to the correct answer. The technology now allows for the possibility of capturing text and transmitting it externally, with AI deriving the correct answer in mere seconds, highlighting the limitations of relying solely on metal detectors or visual checks by proctors.

Consequently, there is a growing call to move away from memorization-based multiple-choice evaluations and to introduce written and essay evaluations that measure critical thinking and problem-solving skills into the CSAT and college admissions process. The National Education Commission is also reviewing plans to gradually expand written and essay evaluations in school assessments as part of its mid- to long-term national education development plan, lending weight to the argument that the evaluation paradigms of school assessments and the CSAT should be linked. With the 2028 college admissions reform plan already indicating a shift to a five-grade system for school assessments and the full implementation of a high school credit system, it is seen as a natural progression for the expansion of written and essay evaluations in school assessments to lead to changes in the CSAT evaluation methods.
Education groups say urgent need for critical thinking assessments in the AI era, while concerns over grading fairness and private education arise
While education circles and various educational organizations agree on the need to move toward an evaluation system that overcomes the limitations of multiple-choice CSATs and aligns with the AI era, there are differing opinions on specific methodologies and the pace of implementation.

Progressive teacher organizations, such as the Korean Teachers and Education Workers Union (KTU), and educational civic groups argue that it is urgent to transform the CSAT into a qualification exam and introduce written and essay evaluations to alleviate excessive competition and rote learning. They contend that since AI has already surpassed humans in simple knowledge memorization and problem-solving techniques, the transition to written and essay evaluations that assess students' critical thinking and creative problem-solving processes cannot be delayed any longer. They emphasize that written and essay evaluations and absolute grading should first be firmly established in high school assessments before the CSAT is also reformed to focus on written and essay formats, which would lead to meaningful changes in classroom instruction.

Conversely, conservative teacher organizations, such as the Korean Federation of Teachers' Associations (KFTA), and some parent groups advocate for caution regarding the hasty introduction of written and essay evaluations. They argue that implementing such questions in a high-stakes national exam taken simultaneously by around 500,000 examinees each year would make it difficult to ensure the objectivity and fairness of grading. Given the sensitivity of South Korean society to a one or two-point difference in college admissions outcomes, there is a significant concern that large-scale disputes over grading criteria and results could arise, potentially exacerbating the pressure on teachers and leading to an increase in private tutoring for essay preparation.
“Switching to written evaluations alone won't stop AI cheating... Independent testing environments and grading verification are priorities”
Education experts warn that discussions on introducing written and essay evaluations should not be limited to merely changing question formats. They stress that enhancing monitoring systems to block external assistance during exams and establishing strict grading criteria must go hand in hand. With generative AI capable of producing plausible essay responses and solutions, if security in examination venues is compromised, written evaluations will not be immune to AI cheating.

Moon Hyung-nam, president of the Korean AI Education Association and a professor at Sookmyung Women's University, stated, "The belief that simply changing from multiple-choice to written or oral formats will resolve AI cheating is a misconception." He pointed out that since generative AI can also write essay responses, the effectiveness of cheating prevention ultimately depends on how well unauthorized external assistance is blocked and how an environment is created for examinees to independently produce their answers.

Moon further emphasized that to expand written evaluations, it is essential to thoroughly verify the consistency of grading criteria and among graders, along with designing questions that ask for evidence supporting conclusions. In the case of the CSAT, which has a large number of examinees, it is crucial to consistently apply device management and monitoring procedures before the exam. He concluded that when reviewing exam reform plans, two principles must be clearly established: "What abilities are we trying to assess?" and "How can we confirm that examinees demonstrated those abilities independently?"

Another education assessment expert noted, "Given the significant practical constraints of immediately converting the entire CSAT to written and essay formats, it would be more realistic to first stabilize the written and essay evaluations and process-oriented assessments being considered by the National Education Commission, and then gradually introduce them in certain areas of the CSAT, while also building infrastructure such as AI-assisted grading systems."




* This article has been translated by AI.