Four smart glasses cheating cases in South Korean school exams. The decisive clues were a hand touching the frame and a green light
← Reviews and readings

Opinion

Four smart glasses cheating cases in South Korean school exams. The decisive clues were a hand touching the frame and a green light

South Korea's Ministry of Education surveyed every middle and high school in the country and announced four cases of cheating with AI smart glasses in school exams. The decisive clues were a hand repeatedly touching the frame, a green light, and glasses the student did not normally wear. This article explains why clues that worked at schools, where teachers know their students, cannot be relied on in large-scale exams.

Naoki ·

Read the Japanese original.

Four cases of cheating with AI smart glasses were found in South Korean school exams. The figures come from a full survey of every middle and high school in the country, published by the Ministry of Education on November 9. What the four cases share is that the decisive factor was not detection equipment but the eyes of the supervising teachers. What did the teachers look at? The clues were specific.

Three clues: a hand touching the frame, a green light, and glasses the student did not normally wear

The first clue was behaviour: repeatedly touching the frame of the glasses. Many smart glasses are operated by taps and swipes on the temple, so calling up an answer shows up on site as a hand touching the frame. In all four cases, the supervising teachers had noticed this abnormal behaviour.

The second clue was light. At a high school in North Jeolla Province, a supervising teacher spotted a green light reflected in a student's glasses and caught the cheating. Light from displays and recording indicators does not exist on ordinary glasses.

The third clue was glasses that were different from usual. In Seoul, a student who did not normally wear glasses arrived at the exam wearing a pair. In Jeju, the method was to wear ordinary glasses before the exam and switch to smart glasses during it; a teacher who had received a tip from another student watched closely and confirmed the switch as it happened.

Clue What it is Where it was reported
A hand touching the frame Operating the glasses by tapping and swiping the temple to call up answers shows up as a hand repeatedly touching the frame In all four cases, the supervising teachers had noticed this abnormal behaviour
Green light Light emitted by display and recording indicators is reflected in the glasses At a high school in North Jeolla Province, a supervising teacher spotted a green light reflected in a student's glasses
Glasses different from usual Taking the exam in glasses not normally worn, and switching glasses during the exam In Seoul, the clue was glasses the student did not normally wear; in Jeju, it was the switch, which a tip from another student brought to light

Schools caught it because teachers know their students

The Ministry of Education explained that school record exams are easier to catch because teachers know their students' usual appearance. In other words, at large exams where candidates and proctors meet for the first time (university entrance exams, national exams, qualification exams), the oddity of 'glasses they do not usually wear' is not available. The remaining clues are the hand touching the frame, the light, and the identification points of the model itself. The less a site can rely on that oddity, the more noticing depends on knowing the points to look at in advance.

And some models give off no clue at all

The four cases were found because clues existed. But not every model has clues. Audio-only models without a camera look no different from ordinary glasses at the lens. Models without a display emit no light. Types that only record and communicate cannot be told apart from ordinary glasses by appearance alone. Identification is extremely difficult for some models. If on-site eyesight cannot settle it, the conversation turns to preparing, separately, a procedure for how to check, record, and escalate doubtful cases, and appraisals that confirm the functions of the model itself.

What to look for can be handed over in advance, through manuals and training

How the frame is operated, light from displays and LEDs, the positions of the camera and charging port, and temple thickness that ordinary glasses do not have. These are points where prior knowledge, not attentiveness on the day, makes the difference. South Korea's Ministry has also reached the stage of compiling the caught cases and on-site cautions into materials and distributing them to schools.

Smart Glasses Lab provides this part for Japan.

  • Identification manual preparation: for each model, its identifying features, the criteria for requesting a check, and the escalation procedure for hard-to-judge cases, documented to match the organizer's rules and operations
  • Identification training for exam invigilators: a 90-minute training session with real devices brought in, where participants touch and confirm the identifying features themselves
  • Device appraisal: we verify the functions of the model in question and state its risk category (A, B, or C) and the grounds in writing

The service is for exam organizers and operating bodies. Fees are quoted individually according to the content. Consultations are accepted even before the content is settled. Contact [email protected] .

Naoki

I'm transmitting the actual experience and usage of smart glasses.

About the author and this site →