Tax Accountant Exam Faces Cheating Allegations Amid Record Scores

The controversy surrounding cheating with artificial intelligence (AI) smart glasses, which has already shaken middle and high school exams as well as university entrance essay tests, has now extended to the national qualification exam for tax accountants. In the 63rd Tax Accountant Qualification Exam held this year, the top scorer, identified as A (34), reportedly missed only one question out of 160 in the first round, leading to widespread suspicions of cheating using advanced devices among the examination community.

According to the tax industry and examination community on November 12, voices demanding clarification regarding the alleged use of AI glasses by A have surged in online student forums and blogs following the recent announcement of the final pass list by the Human Resources Development Service of Korea.

To understand the skepticism surrounding A's score, it is essential to examine the unique structure and difficulty of the tax accountant exam. The exam consists of a first round of multiple-choice questions and a second round of subjective (essay and descriptive) questions. The first round includes four subjects: public finance, introduction to tax law, introduction to accounting, and a choice between commercial law, civil law, or administrative litigation law, totaling 160 questions (40 questions per subject). Candidates must score at least 40 points in each subject and an average of over 60 points across all subjects to pass.

However, candidates face intense time pressure to solve complex calculation problems within the limited time. The first session (public finance and introduction to tax law, 80 questions in 80 minutes) and the second session (introduction to accounting and choice subject, 80 questions in 80 minutes) are notorious for their challenging calculation questions, leading many candidates to adopt a strategy of focusing on certain subjects while sacrificing points in others.

This year, A, who took the exam for the first time and passed both rounds in the same year, achieved an astonishing score of 397.5 out of 400 (an average of 99.37) in the first round held on April 25. He only missed one question in public finance and scored perfect marks in the other three subjects: introduction to tax law, introduction to accounting, and commercial law.

In comparison, the average scores of the top scorers in the previous two exams were significantly lower, with the 61st top scorer averaging 85.63 (tax law 82.5, accounting 75) and the 62nd averaging 81.88 (tax law 72.5, accounting 77.5), making A's score unprecedented.

The decisive factor igniting suspicion was the score distribution chart for the first round of the 63rd Tax Accountant Exam, which candidates obtained through a request for information disclosure from the Human Resources Development Service. This year's first round has been labeled as one of the most difficult in history, with only 3,050 out of 16,856 candidates passing, resulting in a pass rate of just 18.09%. The average score for introduction to tax law was a mere 30.18, with over 74% of candidates failing to achieve the minimum passing score of 40. Introduction to accounting also had a low average of 38.93, with a failure rate of 53.57%.

The score distribution chart revealed that only one candidate scored in the 80s for introduction to tax law, and none scored in the 90s. Yet, A was the only candidate to achieve a perfect score of 100 in that subject. Similarly, only one candidate each scored 100 in introduction to accounting and commercial law, all while A was the only one who missed just one question in public finance.

Suspicion was evident from the start. Two days after the first round, A posted on the online forum DC Inside under the nickname 'Tax Cat,' claiming to have missed only one question. After facing accusations of dishonesty, he later verified his score by posting a video of the results on the national qualification exam portal (Q-Net).

Once the score of 397.5 was confirmed through the information disclosure request, candidates began to argue that even top instructors would struggle to achieve such a score under time constraints, raising concerns about possible answer leaks or the use of AI glasses.

The results of the second round, announced on October 28, further fueled doubts. The second round consists of four subjects, each requiring 90 minutes for essay and descriptive responses. A scored exceptionally high marks in accounting (96 points in part one, 95 in part two) and tax law (83.5 in part two), but his score in tax law part one dropped sharply to 57, resulting in an average of 82.87, still securing the top position.

Examination candidates noted that while the first round's multiple-choice questions and the second round's accounting and tax law part two are based on standardized calculations and memorization of legal texts, tax law part one requires comprehensive application of case law and legal principles to complex factual scenarios, an area where current generative AI struggles. They argued that A's notably low score in tax law part one could indicate the use of AI glasses.

A shared his study routine in a personal essay published on November 11, stating that he graduated from the mechanical engineering department at Seoul National University through a special program for employees while working at a company. After leaving his job to start a business, he prepared for the tax accountant exam for about two years, ultimately achieving top scores on his first attempt.

In his essay, he described waking up at 3:50 a.m. and going to bed at 8 p.m. due to a belief that he lacked talent for studying. He recounted studying desperately, even crying while eating kimbap prepared by his mother during lunch breaks. He attributed his success to techniques such as reverse engineering past exam questions, daily blank reviews, image training, and creating condensed notes.

However, following the publication of his essay, a flood of comments demanding clarification appeared on A's personal blog and community forums. Candidates pointed out that the photos of study materials posted on his blog appeared too pristine for someone who had supposedly read them dozens of times, and questioned how a non-major with no background in economics could miss only one question in public finance.

One candidate who claimed to have passed the first round and taken the second round commented on A's blog, stating that AI glasses had not been a significant issue during the first round in April, and that there was no monitoring for AI glasses during the second round, raising further suspicions. They requested A to disclose his mock exam scores, course enrollment history, and the notes he claimed to have written.

Experts, while cautioning against premature conclusions, agreed that A's scores are unusual. A professor of tax law at a university in Seoul, who has served on exam committees, stated in a phone interview with Aju Business Daily that while it is possible for high scorers to emerge in the first round, achieving only one mistake across three subjects is extremely difficult when considering the typical preparation levels of dedicated students.

The professor noted that while A could indeed be an exceptionally talented candidate, the simultaneous emergence of external factors raises questions, emphasizing the need for the Human Resources Development Service to verify historical statistics and conduct fact-checking.

As of now, there is no objective evidence or official announcement from the Human Resources Development Service regarding A's use of smart glasses during the 63rd Tax Accountant Exam. However, the reason this controversy has escalated beyond a mere incident is due to the detection of multiple cases of cheating involving AI smart glasses in various national qualification exams and university entrance tests this year.

According to data submitted by the Human Resources Development Service to the National Assembly, five cases of cheating involving smart glasses were detected in national technical qualification exams this year (two in gas engineering, and one each in information processing, electrical engineering, and fire safety engineering).

In university entrance exams, candidates wearing or possessing smart glasses were caught during the essay tests at Yonsei University (one case) and the University of Seoul (two cases) last October. Additionally, four cases of AI glasses cheating were confirmed in a nationwide survey of middle and high school internal assessments conducted by the Ministry of Education on November 9.

The issue is that most of the detected cases were revealed through the visual observation of proctors rather than advanced detection equipment, relying on subtle signs such as faint lights or electronic sounds from the glasses, or reports from fellow candidates. If a candidate wore smart glasses that closely resembled regular frames and took the exam naturally, the current monitoring system would have no means to detect them.

In fact, a survey conducted last month by the Korean Federation of Teachers' Associations among 2,716 middle and high school teachers found that 82.3% responded that it is difficult to identify smart glasses under current monitoring procedures, and 64.4% expressed concerns about potential disputes arising from misidentifying them as regular glasses.

In response to the growing concerns, the Ministry of Education plans to assign dedicated personnel to identify smart glasses at all 1,312 exam centers for the 2027 College Scholastic Ability Test scheduled for November 19, and will pursue penalties including disqualification from exams for two years and criminal charges for candidates found wearing smart glasses.

The Human Resources Development Service is also introducing a separate supervisory role of 'AI Inspector' for national qualification exams and is discussing technical countermeasures with the Korea Broadcasting and Telecommunications Commission. Meanwhile, the Ministry of Justice is strengthening guidelines for supervising the bar exam scheduled for January next year. However, there are currently no specialized radio detection devices available for this year's college entrance exam, leaving proctors reliant on their observational skills.

Both the examination community and the education sector emphasize that the controversy surrounding the Tax Accountant Exam could undermine the fairness and credibility of national professional qualification exams. They urge the Human Resources Development Service to clarify the facts based on CCTV footage, proctor records, answer sheets, and exam papers, and to implement effective physical inspection measures to prevent the entry of wearable devices in the future.

Copyright ⓒ Aju Press All rights reserved.