AI-Enabled Smart Glasses Emerge as the Latest Cheating Tool in High-Stakes Exams

A New Generation of Academic Dishonesty

Students across East Asia and beyond are deploying AI-powered smart glasses to cheat in high-stakes examinations, prompting urgent responses from education authorities in South Korea, Taiwan, China, and the United Kingdom. The incidents — including two confirmed cases in South Korea and one in Taiwan during a medical school entrance exam — represent a sharp escalation in technologically assisted academic dishonesty, one that experts warn is far more widespread than reported cases suggest.

Confirmed Incidents Trigger Institutional Responses

In South Korea, two individuals were caught using smart glasses during an English language proficiency exam whose results are routinely used by employers in hiring decisions. Both cases, confirmed last month, mark the country’s first officially reported instances of AI-glasses cheating. South Korea’s college entrance exam administrator has since opened discussions with the Education Ministry and regional education offices, though smart glasses — like all electronic devices — were already prohibited from exam rooms.

In Taiwan, a candidate sitting for a medical school entrance exam was identified after proctors observed the student staring unusually at the test paper. A subsequent inspection revealed the spectacle frames were emitting heat. The university involved is now reviewing its rules and standard operating procedures for AI eyewear during examinations.

China acted pre-emptively. For this month’s edition of the gaokao — the national college entrance examination that draws more than ten million candidates annually — authorities mandated screening of all eyewear before candidates entered examination halls. In the United Kingdom, the head of England’s exam watchdog issued a warning earlier this month that AI glasses and smart earpieces risk materially worsening cheating conditions.

How the Technology Works

The mechanics are straightforward and difficult to counter. Students use glasses equipped with embedded cameras to scan examination questions; the device transmits those questions to a connected large language model, which generates answers and returns them as audio feedback or as subtle visual overlays on the lens — all within seconds. The hardware is increasingly slim, thermally inconspicuous, and in many cases indistinguishable from standard prescription frames.

Researcher Meng Zili, an assistant professor at the Hong Kong University of Science and Technology (HKUST), demonstrated the viability of this approach in a controlled experiment. Inspired by spotting a stylish pair of glasses on a student during a proctored exam — which turned out to be ordinary — Meng subsequently tested commercial AI glasses in an undergraduate electrical engineering course. The device, operating by simply scanning the exam paper, scored in the top five of a class of more than 100 students, well above the class average of 72.

“After conducting the experiment, it really raises the question of how much knowledge students actually need to memorize for exams, versus whether we should allow them to use AI during assessment, given what AI is now capable of,” Meng said. His co-researcher, HKUST electrical engineering professor Zhang Jun, noted that the pace of AI development has left frontline educators struggling to keep up — “every teacher feels that,” he said.

A Market Maturing Rapidly

The proliferation of consumer-grade AI wearables has materially lowered the barrier to this form of cheating. Meta launched its first AI-enabled glasses in partnership with Ray-Ban in late 2023 and has since released multiple updated versions; more than seven million pairs were sold globally in 2024 alone. Devices are becoming slimmer, more capable, and cheaper, with some students now renting them to peers — creating a nascent underground economy built on academic exploitation.

Thomas Corbin, a lecturer at Deakin University in Australia who researches AI-powered devices in academic assessment, offered a blunt assessment: “Wearable AI is as much of a challenge to exams as ChatGPT was to essays in 2022 and I just don’t think there is any real way that we can reliably have exam practices moving forward.” He added that confirmed incidents represent only a fraction of actual cases: “If we’re seeing a few cases being reported, we’re seeing a lot more cases not being reported.”

Institutional Responses Remain Piecemeal

Experts argue these responses are insufficient. Kim Myuhng-joo, executive director of the AI Safety Research Institute and a professor at Seoul Women’s University, called for systemic alternatives: “We need educational alternatives — such as having students sign a voluntary pledge acknowledging penalties for possessing smart glasses, and strengthening punishments.”

A Structural Challenge to Assessment

The deeper question raised by researchers is not merely one of enforcement but of educational design. If a commercially available pair of glasses can place in the top five percent of an undergraduate engineering exam, the premise of closed-book, memory-based assessment comes under fundamental scrutiny. The pressure is especially acute in East Asian education systems, where a single examination result can determine university placement, employment prospects, and social mobility for years to come.

The technological trajectory is clear: AI wearables will become more capable, more affordable, and harder to detect. Whether education systems — built on institutional inertia and examination traditions decades old — can adapt at comparable speed remains an open and consequential question.