[Editorial] AI Glasses Spark 'End of Exams'... We Must Overhaul Outdated Testing and Evaluation Frameworks


The evolution of technology is shaking the very foundations of the 'grammar of exams' that has persisted for over half a century. Amid an unprecedented wave of tension surrounding the misuse of AI (artificial intelligence) smart glasses, the 2027 College Scholastic Ability Test was held simultaneously at 1,312 test centers nationwide. The scene, where the long-awaited efforts of students were engulfed in anxiety and confusion over the monitoring of advanced devices, starkly illustrates the crisis facing our educational evaluation system.

The education authorities deployed over 8,400 dedicated monitoring personnel this year and took comprehensive measures, including thorough checks of glasses and worn items during each session, securing hallway bags, and enhancing metal detector searches. However, the tension inside and outside the test centers remains palpable. Students, who should be fully focused on the exam, are resorting to wearing unfamiliar contact lenses and even removing metal buttons from their coats to avoid unnecessary misunderstandings.

It is the government's duty to impose strict penalties on cheating that undermines the fairness of exams and to strengthen management and supervision. However, merely engaging in a 'battle of swords and shields' that forces students into excessive self-censorship and intimidation does not address the essence of the problem. There are clear limitations to completely blocking wearable devices, which are becoming increasingly miniaturized and high-performance, making them difficult to distinguish from ordinary glasses, through on-site inspections and post-punishment measures. This incident serves as a serious warning that the objective and memorization-based national examination system has reached structural limitations in the face of a monumental shift in technological civilization.

A more fundamental question is what and why we are assessing for future generations. In an era where generative AI can instantly search and summarize vast amounts of knowledge and solve standardized multiple-choice questions faster and more accurately than humans, exams that measure rote memorization of correct answers have already outlived their usefulness. The evaluation paradigm that lined up talents based on the accumulation of knowledge and problem-solving skills during the industrial age is effectively signaling the 'end of exams.' We must fundamentally reshape education and admissions to verify how creatively and critically problems can be solved using AI, rather than viewing it solely as a tool for cheating that must be isolated from the testing environment.

Now, the focus of our society's evaluation must boldly shift from 'what do you know?' to 'what can you achieve?' The government and the education sector must immediately activate a practical roadmap for institutional reform that goes beyond mere declarative slogans, starting with reducing written exams that ask for fragmented knowledge memorization in high school credit systems, and significantly expanding oral, essay-type evaluations and project-based assessments to verify process-oriented thinking skills.

Furthermore, the introduction of digital and AI-based evaluations across the entire national talent selection system, including the CSAT, civil service exams, and national qualification exams, must commence, along with the development of new evaluation items that measure AI utilization ethics and problem-solving capabilities. We cannot prepare for the future with makeshift monitoring measures. It is time to engage in social discourse and action to fundamentally redesign the framework for selecting talents that will sustain the nation's long-term development.




* This article has been translated by AI.