In the exam-oriented East Asian society, students are always looking for breakthroughs in order to stand out in the fierce competition. From peeking at classmates' test papers in class to writing answers on their palms or cheat sheets, various cheating methods are emerging in an endless stream. Nowadays, with the development of technology and the pressure of scoring, some candidates have begun to use smart glasses equipped with artificial intelligence (AI) to quietly gain an "advantage" in the examination room.In East Asia, which is regarded as the "center of exam-oriented culture", a battle is taking place regarding the fairness of exams and the ability of the education system to respond.
In the past month, South Korea discovered two cases of candidates using smart glasses to cheat in English proficiency tests. The results of such tests often directly affect job recruitment and career development, and are regarded as one of the key nodes in life. At the same time, during the entrance examination of a top medical school in Taiwan, a candidate caught the attention of the invigilator because he stared at the test paper with abnormal eyes for a long time in the examination room. After inspection, it was found that the frame of the glasses he was wearing was abnormally hot, and it was eventually discovered that it was a smart device. Education authorities and universities have realized that this is no longer an isolated incident, but a precursor to a new type of risk.
China has upgraded security inspection requirements in advance for this year’s national college entrance examination, and uniformly screened the glasses worn by all candidates to prevent the possibility of hidden devices in smart glasses. In the UK, the head of England's examination regulatory agency publicly warned that high-tech equipment such as AI glasses and concealed smart headphones may exacerbate the problem of cheating in exams and threaten the effectiveness of the existing invigilation system. There is widespread concern in the industry that once an "arms race" forms between regulation and technology, it will be difficult for traditional examination scenarios to maintain the original fair baseline.
South Korea stated that the two previously investigated cases of smart glasses cheating were the first related cases officially reported in the country. The authority responsible for university entrance examinations told the media that they are working with the Ministry of Education and local education administration departments to study further preventive measures, although all electronic devices, including glasses, have long been listed as prohibited items in examination rooms. In Taiwan, the universities involved have also launched a comprehensive review of examination regulations and standard operating procedures, especially the inspection specifications and penalty standards for new equipment such as AI glasses.
However, experts caution that individual cases are just the tip of the iceberg, and the real risk lies in hidden cheating that has not yet been discovered or reported. Thomas Corbin, a lecturer at Deakin University in Australia, pointed out that if sporadic cases have been caught, it may mean that there are more behaviors that are hidden from statistics. In his view, this is not only a management issue for individual countries or schools, but a common challenge faced by the global education evaluation system in the context of technological proliferation.
With the rapid iteration of AI technology, smart glasses are becoming thinner and more invisible. At the same time, the integrated AI model can run locally and maintain network connections, making it increasingly powerful and difficult to recognize by the naked eye. This not only makes invigilation significantly more difficult, but also raises broader concerns about privacy, such as covertly filming test papers, recording examination room footage, and even quietly collecting personal information in daily life. As wearable devices gradually move from "novelty gadgets" to mass consumer products, the tension between technological development and institutional norms has become increasingly obvious.
Since Meta, an American technology company, launched its first AI smart glasses in cooperation with Ray-Ban in 2023, it has successively released multiple generations of new products, promoting wearable AI to enter the daily lives of more ordinary users. According to reports, last year alone, sales of related products exceeded 7 million units, indicating that market acceptance and popularity are rising rapidly. In this context, Corbin believes that the impact of AI wearable devices on exams is almost comparable to the disruptive impact of generative AI such as ChatGPT on written assignments in 2022.
“In the face of wearable AI technology, it is difficult to give a definite answer as to whether closed-book examinations in the traditional sense can still be maintained in the future,” Corbyn pointed out. He believes that if the education sector still relies entirely on the past proctoring model and anti-cheating methods, it is likely to gradually lose control of the assessment process. For universities and employers who urgently need to identify talents through examinations, this uncertainty will directly impact the credibility of admissions and employment standards.
At the Hong Kong University of Science and Technology, a student wore a pair of fashionable glasses during an exam, which caught the attention of the invigilator Meng Zili. As a scholar who has studied AI glasses for a long time, he initially suspected that they were a pair of smart devices, but eventually confirmed that they were just ordinary glasses, but this episode prompted him to conduct a more systematic experiment. He decided to introduce commercially available AI glasses for actual testing in the final exam of an undergraduate-level electrical engineering course to test its feasibility in a real exam environment.
The experimental results were unexpectedly "eye-catching". The content of the test paper is captured through the camera, and the AI glasses transmit the questions to the connected large language model in real time. The model then generates answers and displays the answers on the lenses. The entire process can be completed almost unnoticed. According to test data, the scores generated by this system ranked among the top five in a class of more than 100 students, much higher than the class average score of 72 points.
"This experiment forces us to rethink: How much knowledge do students need to remember under the current examination model? Should we consider conditionally allowing the use of AI tools in assessment?" Meng Zili said. He emphasized that as AI capabilities continue to improve, educators must reflect on the real purpose of exams, whether it is to test memory and skills, or to test problem-solving and critical thinking skills.
Zhang Jun, another person in charge of the project and a professor of electrical engineering at the Hong Kong University of Science and Technology, pointed out that the development speed of technology and AI is much faster than the pace of school teaching and examination system reform, which is a common source of pressure for frontline teachers. “The question we really need to answer is how quickly the education system can redesign teaching methods and assessment mechanisms to adapt to this wave of technology,” he said. In his view, if the old framework continues to be used, students, teachers and the examination system will be forced to be passive and defensive.
Kong Zhaoxiang, director of the Center for Artificial Intelligence and Digital Competence Education at the Education University of Hong Kong, proposed from another perspective that the use of technology cannot be banned simply because of the impact of AI, but it must be guided within reasonable boundaries to become a tool to improve learning efficiency. He believes that the education system should focus on cultivating students' thinking and metacognitive abilities, helping them maintain independent judgment while using AI, rather than relying entirely on machine output. "We should use technology and we should use AI, but the key is not to completely outsource our thinking ability to it," Kong said.
In societies like East Asia, where test scores are highly valued, a key test is often related to future studies, career and even social status. Parents and students naturally tend to test the edges of the rules. However, when devices such as AI smart glasses begin to enter the examination room in a concealed, portable and relatively low-cost manner, the invigilation system originally based on human eye observation and simple metal detection is obviously unable to shoulder the heavy responsibility of maintaining fairness alone. From China to South Korea, from Taiwan to the UK, education and regulatory agencies are trying to find a new balance: on the one hand, they must maintain the credibility of exams, on the other hand, they cannot completely negate the positive potential of technology in teaching and learning.
It is foreseeable that the controversy surrounding AI smart glasses will not stop at the examination room, but will also extend to workplace assessment, professional qualification certification, and other scenarios that rely on examination selection. In this process, society needs to constantly adjust the boundaries between "efficiency" and "justice," "innovation" and "norms." For those students who are preparing for key exams, how to maintain integrity while utilizing technology will also become an unavoidable "must answer" question as they move into the future.