South Korea Tax Exam Top Scorer Admits Using Generative AI

South Korea's top tax accounting exam scorer admitted to using a microcamera, smartphone and generative AI during two tests. Police received his written

Quick Answer

South Korea’s top scorer in the national exam for aspiring tax accountants admitted to using generative AI during two tests after earning 397.5 out of 400 points in the first round. Police said they received his written confession and handed the case to investigators.

What Happened in the Tax Accountant Exam?

The candidate, identified only by the surname Han, ranked first in both rounds of testing in April and July. His first-round result amounted to missing only one question out of 160.

Person writes at a classroom desk while holding a phone

Han initially denied online rumors that he had entered the exam hall wearing smart glasses to cheat. He later explained in a video that he used a tiny microcamera and a smartphone, photographed the questions, sent them to an AI tool and checked the answers on his phone.

Han said in the video: “I cheated using electronic devices and generative AI. My passing the exam with the top score was not the result of fair effort.”

He also said that he submitted a written confession to the Seoul police through his lawyer. According to the report, police confirmed that they received the confession and handed the case to investigators.

LEET Perfect-Score AI Allegations (2026)

Lee Jin-woo, who achieved the first perfect score in the LEET law school entrance exam, faces allegations of using generative AI to cheat. Suspicious answer changes and irregularities in exam papers have fueled concerns about AI misuse undermining exam integrity.

Investigations are ongoing, but no confirmed AI-related cheating has been established.

Why Is the Scandal Significant?

The case creates an immediate exam-integrity problem in South Korea as the national college entrance exam approaches. Exam administrators and government officials are considering safeguards ahead of the scheduled Thursday test.

Prime Minister Han Seong-sook said in a Friday statement that the government would closely oversee the process so the exam runs fairly. A senior presidential office official said the presidential chief of staff for policy had called an emergency meeting to put countermeasures in place before Thursday.

Potential cheating has also been reported around the certified public accountant exam, adding pressure on officials to tighten controls across high-stakes testing.

What Does the Evidence Show About AI-Assisted Cheating?

A survey of more than 95,000 undergraduates at 20 research-intensive public universities found that about two-thirds of respondents used generative AI and almost 40% used it monthly or more frequently. At least 9% of students who used AI reported using it to cheat, according to the study published in Science and reported by UC Berkeley.

Person working on laptop while holding paper in library

The survey, conducted in spring 2024 through Berkeley’s Student Experience in the Research University Consortium, found 26% of daily AI users reported cheating with it, compared with 7% of monthly users. The researchers cautioned that banning generative AI may not stop cheating and could hinder students preparing for workplaces that expect AI proficiency.

A separate Stanford SCALE study examined 52 first-year undergraduates in an Introduction to Programming course at a public university in Türkiye. Logistic regression predicted AI-assisted cheating risk with 73.1% accuracy, 0.718 balanced accuracy and an area under the curve of 0.763, based on learning management system activity from the first eight weeks of the semester.

In that study, 23 of the 52 students, or 44.2%, were categorized as high-risk during a proctored final exam. Course-module views, assignment submissions and video-access days were the most consistently selected predictors. The researchers said the predictions are intended to support timely academic guidance, not to establish misconduct or initiate disciplinary action.

How Is Academic Cheating Being Defined?

A qualitative study of postgraduate students’ reflective essays identified two main forms of AI-assisted cheating: using AI to complete academic tasks and improperly using AI-generated content. Ten essays produced 159 meaning units, 34 codes, 12 categories and six themes.

The students linked these behaviors to work pressure, ethical ambiguity, AI capabilities and gaps in institutional policies. Their proposed solutions included clearer guidelines, improved assessment design and stronger ethics education. The study was published in Frontiers in Computer Science.

Key Considerations

  • The tax examination case involved electronic devices and generative AI, according to Han’s admission and the police confirmation described in the report.
  • UC Berkeley reported that AI policies differ by course, ranging from allowing AI throughout exams to banning it completely.
  • The Stanford study states that its risk predictions are intended for academic guidance rather than disciplinary action.
  • The postgraduate study identifies unclear institutional policies and ethical ambiguity as factors students associated with AI-assisted cheating.

What to Do Now

  1. Review the AI rules that apply to each course, assessment and professional examination you take.
  2. Check whether electronic devices, photographs, external tools and generative AI are prohibited.
  3. Use learning management system activity to identify students who may need academic support.
  4. Create assessments that test knowledge and skills through work that cannot be completed with AI assistance.
  5. Provide clear guidance on acceptable AI use and acceptable academic assistance.

FAQ

What Score Did the South Korean Tax Exam Top Scorer Earn?

Han scored 397.5 out of 400 points in the first round, which was described as missing only one question out of 160.

How Did Han Allegedly Use AI During the Exam?

Han said he used a tiny microcamera and a smartphone to photograph exam questions, sent the images to an AI tool and checked the answers on his phone.

What Are the Police Investigating in the Tax Exam Case?

Police said they received Han’s written confession through his lawyer and handed the case to investigators.

What Did the Survey of More Than 95,000 Undergraduates Find?

About two-thirds reported using generative AI, almost 40% used it monthly or more frequently, and at least 9% of AI users reported using it to cheat.

Can Early Learning Activity Identify AI-assisted Cheating Risk?

A Stanford SCALE study found that logistic regression predicted risk with 73.1% accuracy from the first eight weeks of learning management system traces. The researchers said the model is for academic guidance, not proof of misconduct.

This article reports on publicly available information from the sources listed below. We have no reason to believe it is inaccurate, but we have not independently verified every claim, and details may change as the story develops. This is not financial, legal or investment advice.

Sources


See our editorial disclaimer for how our articles are produced.

🤖
Is your business visible to AI assistants?

Run a free scan to see your AI Visibility Score, SEO rating, and local citation accuracy.

Check Your Score →