In the heart of one of the most sensitive processes of the Argentine health system – the entrance to medical residences – an unexpected crack has opened. More than 7000 health professionals paid Single examination of medical residences. A month later, the Ministry of Health of the Nation detected an unprecedented statistical phenomenon: a wave of extraordinarily high scores, mostly from applicants trained in foreign universities.
The suspicion quickly resulted in action: the suspension of the order of merit and the call to a new evaluationofficialized through Resolution 2274/2025 published in the Official Gazette. But what at first seemed an academic validation problem was silent, silently, an investigation into the possible use of artificial intelligence, spy devices and remote assistance technology during the test.
How fraud was perpetrated
The possibility that technology has been used to commit academic fraud in a national exam is not science fiction: it is a plausible and increasingly real hypothesis in high competitiveness and imperfect supervision contexts. In particular, the use of Smart lenses with hidden camera and connection to AIcombined with attendees such as Chatgpt, today represents one of the most sophisticated roads to evade controls and obtain answers in real time without being detected.
These devices, which are already available in the Asian and European market-like the Ray-Ban Meta Smart Glasses or the Rokid Max Pro— They integrate an almost invisible front microcamara, Wi-Fi or 5G connection, and in some cases, projected augmented reality screens directly on the lens. In a face -to -face exam, with human personnel watching and without active electronic detectors, these lenses can easily go unnoticed.
The mechanism is simple and efficient: the applicant enters the classroom with the lenses on. At the beginning of the evaluation, the camera is discreetly active to capture the questions, either on paper or on screen. The image is transmitted in real time to a cloud server where an AI, as a chatgpt integrated through an API, analyzes the slogan, consults its medical knowledge base and returns the appropriate response. This answer can be projected directly on the lens by means of an HUD screen (Head-Up Display), allowing the user to read it without looking at a cell phone or making evident movements.
Another variant even more accessible and difficult to track is to use a Bluetooth microauricular hidden in the ear and a smart phone hidden in clothes. Through very short voice commands, the applicant can read the questions or parts of them, and receive in seconds the audited answer by an intelligent assistant who can even be on the other side of the world. In both cases, the key to success is in the miniaturization of hardware and the processing rate of AI, which today allows to solve complex clinical questions in less than three seconds.
It is not the first time that a massive exam is under suspicion. Already during the Covid-19 pandemic, multiple universities and state agencies faced complaints of virtual exams in which control systems They failed to detect trapswhich forced to implement surveillance tools such as ProCotorio or Respondus. However, most of these systems do not apply in face -to -face contexts, where trust falls exclusively on human observation and, in the best case, in the subsequent collection of statistical data.
In the case of the medical residences exam, The use of artificial intelligence could explain why a group of applicants with medium academic trajectories, or even below the average, managed to overcome the threshold of the 90 pointssomething that has historically been reserved for graduates with the best averages of national public universities.
A state problem
The alarm was lit a few days after July 1, when the single exam was taken in venues from all over the country. According to data revealed by the Ministry of Health, the 33.6 % increase in the highest notes Regarding previous editions, he had no correlation with the academic averages of the candidates or with the historical performance of their universities. It was the presidential spokesman, Manuel Adorniwho publicly confirmed that more than 200 cases were being reviewed for “suspicious results.” The first projections showed that little recognized foreign universities had placed dozens of graduates among the 500 best places in the ranking, something unprecedented.
The decision was immediate: the order of merit was suspended, the schedule was reprogrammed and two new evaluation instances were established: a oralfor those who demonstrate academic consistency, on August 4 and 5; and another writtenscheduled for August 7, with 100 multiple choice questions. Those who do not present will be directly excluded from the process.
The reaction was also immediate in social networks and between groups of young doctors. Some pointed out that the resolution was discriminatory towards foreign graduates, while others denounced that the original exam control system had been lax and without adequate technological elements. No signal inhibitors were used or bodily scan was performed on applicants. Nor were there surveillance cameras or video record of the instance. All that allowed – according to sources close to the prosecutors involved – that a line of research focused on the use of spy technology is opened, with collaboration orders for smart lens suppliers.
The Ministry of Health, conducted by Mario Lugoneshas declared that the results of the new evaluation will be final, and that, in no case, the applicants may obtain a higher rating than that originally obtained. This decision seeks to discourage any incentive to repeat fraudulent practices or seek mechanisms to artificially improve performance.
The news trial
By capturing the image of one of the pages obtained from the video that already circulates in the networks, engraving of a presumably Ecuadorian student, News magazine analyzed the image with chatgpt to obtain the answers using the following prompt: “Analyze the questions of this list as a specialist in Clinical Medicine and suggest an answer between the multiplectic options to each of those statements.” The answer in seconds was the following and the method could have been carried out simply for each of the pages.
Question 1:
What is the first image diagnosis study that should be indicated in a 7 -month -old child with febril urinary suspicion?
Options:
a) Cistouretrography
b) Magnetic resonance
c) renal scintigraphy
d) I radiography with a vionze plate
e) Renal and urinary tract ultrasound
Analysis:
In a child with suspicion of feverish urinary infection, the first election study is renal ultrasound and urinary tract, since it is non -invasive, detects malformations and is the initial step before more complex studies.
Answer: e) Renal and urinary tract ultrasound.
Question 2:
What characteristic corresponds to the definition of a “rate” as a health indicator?
Options:
a) It is used as a qualitative indicator
b) results from the division of two numbers that are part of different universes
c) relates the population at risk to a certain event with the non -exposed population
d) It is expressed as quotients in relation to a denominator
Analysis:
The definition of fee in epidemiology is the relationship between the number of events (numerator) and the population at risk (denominator) in a certain time.
Answer: d) It is expressed as quotients in relation to a denominator.
Question 3:
What is the definition of chronic renal disease (ERC)?
Options:
a) FG less than or equal to 60 ml/min and/or presence of albuminuria greater than 30 mg for more than 3 months
b) FG less than or equal to 60 ml/min and presence of albuminuria greater than 30 mg for more than 3 months of evolution
c) FG less than 60 ml/min and presence of albuminuria greater than 30 mg for more than 6 months
d) FG less than or equal to 60 ml/min and presence of albuminuria greater than 30 mg for more than 6 months
Analysis:
The current definition (Kdigo) establishes: FG
Answer: a) FG ≤60 ml/min and/or albuminuria> 30 mg for more than 3 months.
Question 4:
What recommendation would you make about the realization of a colonoscopy to Marcos, 38, which presents as a history the death of a first -degree relative due to colorectal cancer at 45 years?
Options:
a) Start with colonoscopy at age 50 and repeat every 5 years
b) Colonoscopy every 10 years at 40 years
c) Colonoscopy at 40 and repeat every 5 years
d) Colonoscopía at 40 and repeat every 10 years
Analysis:
In first -degree relatives with CCR before the age of 60, the recommendation is to start at age 40 or 10 years before the index case, which happens first, and repeat every 5 years.
Answer: c) Colonoscopy at 40 and repeat every 5 years.
Question 5:
50 -year -old patient has type 2 diabetes in metformin treatment 2,550 mg/day, which persists with HBA1C of 8% and high glycemia. What is the most appropriate therapeutic behavior?
Options:
a) Maintain dose and reassess
b) Increase dose of metformin to 3 g/day
c) Associate another hypoglycemic drug (insulin or oral antidiabetic) and reinforce lifestyle changes
d) Increase dose of metformin to 1,000 mg every 8 h
Analysis:
The patient is already with maximum doses of metformin (> 2,000-2,550 mg/day). If bad control persists, the behavior is to add another agent (insulin or oral antidiabetic) together with lifestyle measures.
Answer: c) Associate another hypoglycemic drug and reinforce lifestyle changes.
Suggested final answers:
and)
d)
to)
c)
c)
You may also interest you
By rn
Image gallery
