AJP Asia Intelligence Agency

AI Glasses Expose Cheating in Tax Accountant Exam

A shocking revelation has emerged from the 63rd Tax Accountant Qualification Examination, where a candidate achieved an extraordinary score by answering all but one of the 160 questions correctly. This individual has confessed to cheating using a miniature camera and generative artificial intelligence (AI). The myth of the 'top scorer' who received perfect scores in three subjects, including the notoriously difficult Introduction to Tax Law, has ended in a bold fraud scheme that allowed them to pass not only the tax accountant exams but also the first stage of the Certified Public Accountant (CPA) exam.


What is particularly alarming is that the monitoring system at the examination venues was completely ineffective during the three national qualification tests. Had the candidate not flaunted their top score online, the situation might have remained undiscovered. Kang Seong-tae, known as the 'God of Study,' recently revealed in a YouTube video that one reason he decided to stop his activities in 2022 was due to the emergence of AI glasses. He stated, "While students spend an average of over 100 million won on private education over 12 years to answer a killer question, an elementary student wearing AI glasses can solve problems better." He added, "If one case has been detected, there are dozens more that have not, and true cheaters never aim for the top but rather pass without drawing attention."


With the 2027 College Scholastic Ability Test just four days away, government officials are still focused on physical monitoring measures. The Ministry of Education has warned that it will deploy identification personnel at examination sites and pursue disqualification and criminal penalties for those caught cheating. However, these measures seem inadequate compared to the rapid pace of technological advancement. As Kang pointed out, in places like Shenzhen, China, it is possible to customize AI-equipped ultra-thin metal or rimless glasses in just one hour, and they can be rented for as little as 10,000 won on exam day.


Moreover, simply banning glasses will not resolve the issue. Lim Hyun-seo, a lawyer running a legal tech firm, commented on the situation, saying, "It doesn't even have to be glasses. Shirt buttons, collars, rings, fingernails, as well as homemade wearables hidden in contact lenses or ear canals are becoming a reality. The current monitoring system cannot keep up with the advancing technology." He emphasized that relying on analog methods to catch faint lights reflecting off lenses is bound to fail.


The real question posed by the 'Tax Accountant AI Glasses Gate' is, "Why are we still asking humans to solve exams that can yield perfect answers in seconds when photographed and sent to AI?" Lim's assertion that society and the economy have become an 'item battle' that cannot function without AI raises a critical point: why are we insisting on a 'no-item' exam for professionals who will never work without tools? Kang echoed this sentiment, questioning, "If someone wearing AI glasses can provide smarter answers than someone who has studied for 20 years, who should companies hire?" He lamented that while the world has changed dramatically, educational policies remain stuck in the past.


Prestigious universities abroad have already abandoned the reckless game of 'AI monitoring' and are experimenting with overhauling their assessment systems. Recently, leading engineering schools and academic institutions, including the Massachusetts Institute of Technology (MIT), have recognized that traditional closed-book exams and simple coding tasks no longer effectively differentiate student capabilities. They are now discussing a comprehensive redesign of examination systems. Instead of outright banning AI use in testing, they are considering allowing the use of AI tools while requiring critical evaluation of AI-generated results, or shifting the focus to one-on-one in-depth interviews with professors and real-time problem-solving assessments on whiteboards.


In fact, the recent cheater in the tax accountant exam scored nearly perfect in the first stage's multiple-choice questions (99.37 points) and in the second stage's accounting (95-96 points), but struggled with the second stage's 'Tax Law Part 1,' where complex legal reasoning was required, scoring only 57 points. This paradoxically demonstrates the limitations of AI and the unique areas where human evaluation is essential.


AI glasses are not merely a cheating tool; they signify the death knell for decades of rote memorization and objective assessment systems. Increasing the number of detection devices and intensifying monitoring alone will not preserve the legitimacy and fairness of existing examination systems. The era of requiring students to memorize knowledge that AI can process in seconds, while searching for electronic devices during exams, has come to an end. Educational authorities and national examination agencies must no longer ignore the imperative to fundamentally change the essence of assessment.





* This article has been translated by AI.