The Machine That Won't Say Why

US · CanadaAmérica Latina

Weekly AI Watch

Every week, what changed and what it means for you. No hurry, no alarm.

Week of 30 October 2026EN | ESPast editions15,230 papers found · 14,455 read · 4,301 belonged · 46 stories passed every check

Observatorio de la
Inteligencia Artificial

Cada semana, qué cambió y qué significa para usted. Sin prisa y sin alarma.

Semana del 30 de octubre de 2026EN | ESEdiciones anteriores15,230 estudios encontrados · 14,455 leídos · 4,301 correspondían · 46 notas pasaron todas las comprobaciones

weeklyAI · Week of 30 October 2026weeklyAI · Semana del 30 de octubre de 2026

← Your job← Su trabajo

interviews · The journal of nursing research : JNR · la publicación, 1 oct 2026 · gratis

A una enfermera la evaluó una máquina y nadie le dijo cómo la calificó

En cinco grupos de conversación en Corea del Sur, 10 postulantes a enfermería y 10 formadores hablaron de las pruebas con inteligencia artificial. Son opiniones de 20 personas, no una medición de a quién contratan.

Versión breve · la versión detallada sigue, unos 7 min

Pregunte al chat de weeklyAI

Pregúnteme sobre la IA en su vida: le señalo las notas que weeklyAI ha publicado sobre el tema, esta semana y antes.

Las conversaciones se guardan mientras exista weeklyAI, para mejorar la publicación. Se responde en el idioma en que usted escribe.

El estudio, de un vistazo
Quiénes: Aspirantes a enfermería y educadoras de enfermería (profesoras y jefas de personal)
Cuántos: 10 aspirantes y 10 educadoras, en cinco grupos de conversación
Dónde: Corea del Sur
Cuándo: Entre febrero y agosto de 2024
Tipo de estudio: interviews
Quién lo hizo: Investigadores de universidades de Corea del Sur (Gachon, Konkuk y Kyung Hee)
El límite que importa: Son 20 personas de un solo país, con experiencia reciente; no mide a quién contratan.

Cuando una máquina lo filtre para un empleo, pregunte qué criterios usa y qué revisa una persona después.
Lectura de weeklyAI

Ver · 0:26

"Esto era frustrante porque ni los que pasaban ni los que no pasaban sabían la razón". Lo dijo una postulante a enfermería en Corea del Sur, según un estudio que reunió a 10 postulantes y 10 formadores de enfermería, entre profesores y jefes de servicio, en cinco grupos de conversación entre febrero y agosto de 2024.

El estudio no detalla qué pruebas hizo este grupo. El artículo recoge un estudio de otro equipo que describe el formato general de estos sistemas, según un desarrollador de Corea del Sur: una autopresentación, una entrevista breve en video, un test de personalidad, preguntas de situación, un juego en línea para medir razonamiento y una ronda de preguntas a fondo.

Detrás de esa frase hay tres quejas. Los criterios para calificar no son claros. La entrevista automática no capta el trato de persona a persona que exige la enfermería. Y el resultado puede depender de la red, la luz y el ruido. Una postulante contó que el sistema se trabó durante tres horas por la cantidad de gente conectada. Entró temprano, cuando había menos gente, y buscó un café internet con buena luz y silencio.

También hubo algo a favor. Los jefes de servicio vieron en la prueba un filtro útil para contrataciones numerosas, que ahorra tiempo de entrevista y cansancio del personal. Las postulantes valoraron poder presentarse a cualquier hora y desde cualquier lugar, sin faltar a clases.

El estudio no midió quién fue contratado ni cómo trabajó después. Tampoco analizó si estas pruebas tienen relación con que una enfermera se quede o renuncie, así que no puede decir que retengan o pierdan gente. Y son 20 personas de un solo país, con experiencia reciente en estos sistemas: nada de esto describe cómo filtran las empresas en México, en Estados Unidos o en Canadá.

Lo que los participantes propusieron fue concreto. Que las pruebas se diseñen pensando en la enfermería, con comunicación y empatía. Que sus resultados alimenten la entrevista cara a cara. Que las instituciones publiquen los criterios.

Cuando una máquina lo filtre para un empleo, pregunte qué criterios usa y qué revisa una persona después. En esta investigación, los propios formadores querían que la máquina ayudara a la entrevista y no la sustituyera.

Qué significa para usted

Cuando una máquina lo filtre para un empleo, pregunte qué criterios usa y qué revisa una persona después. En este estudio, los propios formadores querían que la prueba con inteligencia artificial alimentara la entrevista cara a cara y no la sustituyera. Esa fue su opinión, no un resultado medido, y viene de 20 personas en Corea del Sur, así que no describe lo que pasa en México, Estados Unidos ni Canadá.

LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768

Quién pagó: El estudio contó con el apoyo de una subvención de la National Research Foundation of Korea financiada por el gobierno coreano (MSIT; n.º RS-2026-25472700); el artículo no dice nada sobre si los financiadores tuvieron influencia ni sobre equipos o programas prestados a los investigadores.

No tome esto como consejo médico profesional.

Los hallazgos de otros estudios que aquí se mencionan los conocemos por este documento, que fue el que leímos; no abrimos cada uno de esos estudios.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1491 palabras · unos 7 minLeerla →Cerrar

Una máquina filtra a las enfermeras en Corea del Sur; las aspirantes dicen que nadie les explica cómo las califica

Un estudio con 20 personas en Corea del Sur describe cómo se sienten las aspirantes y sus maestras ante la evaluación automática: les gusta la rapidez, pero no saben qué se mide ni por qué pasan o no.

Imagine la escena: una aspirante a enfermera, en su último año de carrera, se sienta frente a una computadora durante casi una hora. Se presenta, contesta preguntas, resuelve un juego en línea. Al final no hay una persona que le estreche la mano, sino un programa que la califica. Ese programa es la evaluación de competencias basada en inteligencia artificial, y en Corea del Sur ya se usa para contratar enfermeras en hospitales grandes. Un grupo de investigadores quiso saber qué piensan de ella las aspirantes y las personas que las preparan.

El estudio se hizo entre febrero y agosto de 2024 en Corea del Sur. Los autores son Hyeongsuk Lee, Hyeongju Ryu y Hye Jin Yoo, y los pagó el gobierno coreano por medio de un fondo nacional de investigación. Reunieron a 20 personas en cinco grupos de conversación: 10 aspirantes con experiencia reciente en estas evaluaciones y 10 educadoras de enfermería, siete profesoras y tres jefas de personal de hospital. No se midió quién fue contratada ni cómo le fue después en el trabajo. Se escucharon opiniones.

De esas conversaciones salieron cuatro grandes temas. El primero: dudas sobre cómo y con qué se califica. Las aspirantes contaron que no sabían si las evaluaba una máquina o una persona, y otra dijo que había muchos rumores sobre los criterios y que tanto las que pasaban como las que no sabían la razón. Otra dijo que aprendió que la enfermería es un trato de persona a persona, y que usar inteligencia artificial para escoger enfermeras le parecía una contradicción. Una profesora lo resumió así: la evaluación tiene un lado mecánico que no deja ver el lado humano del estudiante.

El segundo tema: la eficiencia. Las aspirantes agradecieron poder presentarse a varias convocatorias sin faltar a clases ni viajar lejos. Una jefa de personal contó que la herramienta le quita carga de trabajo y le deja más tiempo para su labor de enfermería. El tercer tema: las dificultades de prepararse. Una aspirante relató que el sistema se cayó durante tres horas porque había demasiada gente conectada, y que terminó buscando un café internet bien iluminado y sin ruido. Otra practicó hablar como robot, porque había oído que así se saca más puntaje. Los educadores temen que quien no tenga buena computadora o buena conexión quede en desventaja.

El cuarto tema: qué habría que mejorar. Los participantes pidieron un programa hecho para la enfermería, que tome en cuenta la comunicación y la empatía, con criterios claros y revisados de vez en cuando. Una jefa de personal dijo que el resultado de la evaluación le serviría como dato útil para la entrevista cara a cara, y las educadoras recomendaron usar la evaluación para apoyar los métodos tradicionales, no para sustituirlos. Los propios autores reconocen los límites: son 20 personas en un solo país, todas con experiencia reciente en este tipo de prueba; no se trata de una medición objetiva, y los resultados pueden quedar viejos pronto porque la tecnología cambia rápido.

Piense en una empresa como un colador grande. La evaluación automática es la primera malla: por ahí pasan cientos de solicitudes y solo algunas llegan a la entrevista personal. Según el propio artículo, estas pruebas suelen incluir una autopresentación, una entrevista breve en video, un test de personalidad, preguntas de situación, un juego en línea que mide razonamiento y capacidad de hacer varias cosas a la vez, y preguntas de fondo¹. Es decir, la máquina no solo ordena papeles: observa cómo se habla, cómo se reacciona y cómo se juega.

El artículo señala que herramientas parecidas se usan en Estados Unidos y Japón desde 2018, y que en Corea del Sur se están adoptando cada vez más². En ese país llevan cuatro años aplicándose a la selección de enfermeras, y aun así, según los propios autores, nadie había examinado en la literatura si de verdad sirven para escoger a las mejores en las competencias que importan³. Lo curioso es que el sistema ya está en la puerta de los hospitales y la evidencia va detrás, a paso lento.

Hay otras investigaciones sobre entrevistas hechas por máquina, aunque solo pudimos leer el resumen. Un estudio con 225 personas en China y 213 en Pakistán encontró que las entrevistas automáticas se asociaron con menos transparencia percibida, peor comunicación y menos ganas de seguir en el proceso, y que el efecto fue más fuerte en Pakistán que en China; la confianza en los sistemas de inteligencia artificial reducía ese malestar⁴. Los autores recogieron los datos con muestreo de conveniencia, es decir, con voluntarios que se ofrecieron⁵.

Otro trabajo, este con experimentos y cuyo texto completo está detrás de una suscripción, describe una paradoja: si la empresa avisa que usa inteligencia artificial para evaluar, eso se lee como transparencia, pero a la vez cambia cómo los evaluadores juzgan a quien la usó para prepararse, y esa persona puede caer mejor o peor según el caso⁶. Y un tercer texto, del que solo leímos el resumen, sostiene que las medidas de justicia por sí solas no bastan para auditar estos sistemas, y propone revisar quién responde cuando algo sale mal y cómo se puede reclamar⁷. Nada de esto se probó en Corea del Sur ni en enfermería; son piezas de un mismo rompecabezas que se está armando en varios países a la vez.

Así lo leemos nosotros. Cuando una decisión que cambia la vida de alguien se toma con un cálculo automático, casi nunca se explica con qué datos se armó ni qué regla se aplicó. La persona solo ve el resultado, no el camino. Eso es lo que aparece en estas conversaciones: la duda no es solo el puntaje, sino el vacío. Y el vacío se llena con rumores y con tácticas inventadas para caerle bien al sistema, mientras quienes aconsejan tampoco saben qué recomendar. Esperamos que en otros lugares pase lo mismo mientras no haya criterios publicados. Usted sabrá que nos equivocamos si un día las instituciones publican sus criterios y las personas pueden explicar por qué pasaron o no.

Conviene hacer una concesión: a las máquinas se les atribuye limpieza y ausencia de favoritismos, y por eso se les entrega no solo lo pesado y repetitivo, sino también lo delicado, para no cargar con la responsabilidad de decidir. Es una razón que tiene su lógica: una entrevista de diez minutos también juzga con poco. Pero vemos el riesgo de que las quejas se dirijan al aparato y no a quien lo contrató y eligió qué se mide. Si te toca una selección con máquina, pide por escrito qué se evalúa y con qué datos; guarda tu propio registro de lo que hiciste y de lo que te respondieron, y compártelo con quienes estén en la misma fila que tú.

Hay un tercer patrón que nos parece el más útil para quien está del otro lado. Quienes viven una reconversión impuesta no solo se resisten: rearman sus lazos y buscan calificarse de otra manera. En estas conversaciones aparecen las redes entre egresados y estudiantes, la práctica imitando al sistema y la petición de que las escuelas enseñen destrezas digitales como parte de la carrera, no como un extra. Esperamos ver cada vez más de eso dondequiera que se implante un filtro opaco. Nos equivocaríamos si las personas se prepararan cada una por su cuenta, sin buscar a otros. Lo que usted puede hacer es buscar a alguien que ya pasó por el mismo filtro y preguntarle qué hizo; si puede, organice con compañeros un espacio para compartir lo aprendido y practicar lo que el sistema sí mide.

El estudio no dice que la máquina escoja mejores o peores enfermeras, porque eso no se midió. Tampoco dice que el filtro haya mejorado la justicia o la eficiencia: esas fueron percepciones de los participantes, no efectos comprobados. Y todo ocurrió en Corea del Sur, con personas que ya habían vivido ese sistema; nada de esto describe cómo se contrata en México, en Colombia o en Estados Unidos. Si en su trabajo o en el de un familiar aparece una evaluación automática, la pregunta útil no es si la máquina es justa en abstracto, sino quién eligió qué se mide y quién firma el resultado.

Volvamos a la aspirante frente a la computadora. Ahora sabemos que el programa que la califica puede ser rápido y hasta cómodo, y que aun así ella no sabe qué vio la máquina en su cara, en su voz o en su juego. La próxima vez que enfrente un filtro así, pregunte: ¿me pueden decir por escrito qué se evalúa y quién responde si se equivoca?

De dónde sale cada dato de contexto, y cuánto leímos de cada documento

LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - el artículo del que trata esta nota — el artículo completo — el pasaje: "these assessments (as specified by a leading developer in South Korea; Midas IT, 2023 ) generally consist of a self-introduction, simple video interview, personality test, situational questions, an online AI game (to evaluate reasoning and multitasking), and in-depth questions."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - el artículo del que trata esta nota — el artículo completo — el pasaje: "AICAs and similar AI-based selection tools have been used in the United States and Japan since 2018, and are increasingly being adopted by institutions in South Korea ( Kim & Kwon, 2022 )."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - el artículo del que trata esta nota — el artículo completo — el pasaje: "despite the use of AICAs in nurse selection in South Korea for 4 years, the effectiveness of AICAs in vetting and selecting nurses based on these competencies in the South Korean context has yet to be examined in the literature."
Zhao G, Liang Q, Alam F, Hossain MB. (2026). AI-based interviews and applicant perceptions: insights from China and Pakistan. Frontiers in Psychology. 10.3389/fpsyg.2026.1886963 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The results show that AI-based interviews were negatively associated with perceived transparency, interactive communication, and job pursuit intention, with stronger negative associations observed among respondents in the Pakistani sample than in the Chinese sample."
Zhao G, Liang Q, Alam F, Hossain MB. (2026). AI-based interviews and applicant perceptions: insights from China and Pakistan. Frontiers in Psychology. 10.3389/fpsyg.2026.1886963 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Data were collected from 225 respondents in China and 213 respondents in Pakistan using a convenience sampling technique."
Thakur K, Singh A, Jena LK, Srimannarayana M. (2026). Performance expectancy of AI and disclosure consequences: A psychological paradox. Acta Psychologica. 10.1016/j.actpsy.2026.107837 — solo el resumen - el texto completo está tras una suscripción — el pasaje: "Study 2 reveals a psychological paradox: while disclosure of Generative AI use signals transparency, it simultaneously alters evaluators' social and competence-based judgments, producing differentiated effects on likeability and hireability compared to non-assisted applicants."
Authors not given by the registry. (2026). FROM ALGORITHMIC OPACITY TO GENDER-RESPONSIVE ACCOUNTABILITY: AN EXPERT-VALIDATED G-XAI AUDIT FRAMEWORK FOR HIRING AND CREDIT DECISIONS. Acta Innovations. https://doi.org/10.62441/actainnovations.vi.2510 The authors could not be read from the registry. — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The evidence synthesis showed that fairness metrics alone are insufficient, while expert validation supported the framework’s relevance and emphasized gender and intersectional analysis, explainability linked to accountability, human oversight, documentation, contestability, and remediation."

No tome esto como consejo médico profesional.

Los hallazgos de otros estudios que aquí se mencionan los conocemos por este documento, que fue el que leímos; no abrimos cada uno de esos estudios.

interviews · The journal of nursing research : JNR · the paper, 1 Oct 2026 · free

The Machine That Won't Say Why

In South Korean focus groups, 10 nursing applicants and 10 nurse educators called AI screening efficient but could not tell what it rewards; their doubts are perceptions, not measured results.

Short version · the longer version follows, about 6 min

Ask the weeklyAI chat

Ask me about AI in your life: I point you to the stories weeklyAI has published on it, this week and before.

Conversations are saved for as long as weeklyAI exists, to improve the publication. Answers come in the language you write in.

The study at a glance
Who: nursing applicants and nurse educators
How many: 10 nursing applicants and 10 nurse educators
Where: South Korea
When: February to August 2024
Kind of study: interviews
Who did it: Gachon University, Konkuk University, Kyung Hee University
The limit that matters: 20 people in one country; no one measured who got hired or how they performed.

If a machine screens you for a job, watch for whether the employer will tell you what it measures and whether a person makes the final call.
weeklyAI's reading

Watch · 0:30

One nursing applicant, screened by a computer program for a hospital job, put the problem in a sentence: "This was frustrating because both those who passed and those who failed did not know the reason."

That sentence sits on top of a wider doubt. In five focus groups held in South Korea in 2024, 10 nursing applicants and 10 nurse educators, seven professors and three nurse managers, described the AI-based competency assessments now used to screen new nurses there. The programs generally include a self-introduction, a short video interview, a personality test, situational questions, an online game that tests reasoning and multitasking, and in-depth questions.

The doubts came first. Applicants said a computer interview does not fit nursing, which they described as person-to-person work. Professors said the method misses the warm human side of students. Managers said their hospital's own mission was not reflected in the questions. Applicants worried that results could turn on the network, the lighting or the noise in the room. Nurse educators worried that results could depend on whether students had good access to computers and training.

The people in the groups also saw real advantages. Managers called the system a useful filter for large-scale hiring, and said it saved interview time and staff fatigue. Applicants liked being able to interview at any time from any place with a connection, instead of choosing between a class and a hospital.

What the study cannot show is whether the machine picks better nurses. The researchers interviewed 20 people in one country, all with recent experience of these systems. No one measured who got hired or how they performed. The study did not analyze whether the screening relates to nurse turnover or adaptation, so it cannot say whether it keeps people in the job.

The participants made their own proposal. They want the AI results to feed the face-to-face interview, giving human evaluators questions to ask, rather than deciding alone. They also want the rules published and the programs rebuilt around nursing skills such as communication and empathy.

If a machine screens you for a job, watch for whether the employer will tell you what it measures and whether a person makes the final call.

What this means for you

If a machine screens you for a hospital job, ask what it measures and whether a person makes the final call. The 20 people in these groups spoke from South Korea, so weigh their doubts as questions worth putting to any employer, not as proof of what happens here.

Who paid: The study was supported by a National Research Foundation of Korea grant funded by the Korean government (MSIT; No. RS-2026-25472700); the article says nothing about funders having a say or about equipment or software lent to the researchers.

Do not take this as professional medical advice.

The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1134 words · about 6 minRead it →Close

A Machine Screens the Nurses. Nobody Told Them How It Scores.

In South Korea, 20 applicants and educators described a hiring test that saved time and left them guessing. Their answers are perceptions, not proof of who gets hired.

A self-introduction, a short video interview, a personality test, situational questions, an online game, and deeper questions. That is roughly what a nursing applicant in South Korea now meets when a hospital screens new nurses. The machine scores. The applicant waits for an answer.

The test is called an AI-based competency assessment. According to the article's description of what a leading South Korean developer offers, it generally includes a self-introduction, a short video interview, a personality test, situational questions, an online game that measures reasoning and multitasking, and deeper questions¹. The study asked 10 nursing applicants and 10 nurse educators what that experience was like. The educators were 7 nursing professors and 3 nurse managers. The work used five small focus groups, three to five people each, and the interviews ran from February to August 2024². This was a qualitative study: people talking about how it felt, not a measurement of who was hired or how they performed later.

The interviews produced eight subthemes under four themes: doubts about the method and criteria, the efficiency of the assessments, the challenges of preparing for them, and suggested improvements. The frustration is concrete. Applicants could not tell what was being scored, on what basis, or even whether a person reviewed their answers³. One applicant put it plainly: there were rumors about the criteria, and both those who passed and those who failed did not know why⁴.

Picture the job being screened for. Nursing professors in the study argued that the assessment misses the warm human side of students⁵, and one applicant called it a contradiction to use AI to select nurses, whose work is human-to-human care⁶. The same tool has a practical side that the participants did not deny. A nurse manager said the pre-interview data let them prepare for face-to-face interviews, eased the burden of evaluation work, and left more time for nursing itself⁷.

How did the machine get there? These tools have been used in the United States and Japan since 2018 and are being adopted increasingly by institutions in South Korea⁸. Yet, according to the study, although South Korea had used them in nurse selection for four years, their effectiveness in judging nursing competencies had not been examined in the literature⁹. Consider what that gap means: four hiring cycles in which a tool shaped who advanced, and the question of whether it measures the right things was still open.

The trade-off runs through the interviews. One applicant liked that the assessment could be taken at any time and place, so a class did not have to be sacrificed for an interview¹⁰. Against that convenience sat the machinery of taking it. One applicant described the system crashing for three hours when too many people connected at once, and said they logged in early in the morning and searched for a well-lit, soundproof internet café to avoid technical trouble¹¹. A professor worried that students who can use information technology freely and those who cannot will differ, and that uneven information support raises fairness questions¹². Same test, different starting lines.

Then there is the performance itself. One applicant, working on the belief that a robot scores consistency, practiced a stiff, uniform way of speaking rather than their own style¹³. Nursing professors said the assessment has a strong mechanical side that makes it hard to show the warm human side of students, and a nurse manager said it is important to explore the essence of nursing and express one's thoughts clearly, but applicants tend to focus on improving skills only.

Here is how we read it. When a decision is made by a machine or a distant rule, the people on the receiving end are seldom told the reasoning. They fill the silence with rumor. They then shape their preparation around what they guess the machine wants, not around what the job actually requires. Expect more of this wherever an automated screen sits between a worker and a job. You would know we are wrong if applicants could obtain written scoring criteria or feedback explaining a pass or a fail, and if their preparation matched real nursing tasks. You can act on this now: when you face an automated screen, ask the employer in writing what is scored and whether feedback is available, and keep a record of what you were told. Compare notes with others who took the same screen. That is how rumor gets separated from fact.

We also read a second pattern, and it belongs to anyone who has ever been interviewed. People manage how they appear to an audience; they keep a private backstage and stage a polished front. When the audience is a camera that scores, the backstage shrinks. Expect applicants to rehearse image rather than describe their actual work. If applicants told us they were judged on unscripted, interactive tasks and their ordinary manner was accepted, the expectation would fail. So notice when you are performing rather than describing, and practice telling real situations you handled. Ask whether the assessment leaves any room for unscripted exchange.

What would have to change? The authors call for algorithms built around core nursing competencies, clear evaluation criteria, and collaboration between nursing and AI experts¹⁴. A nurse manager in the study hoped AI results would serve as useful background for the face-to-face interview rather than a verdict¹⁵. That is a supporting role, not a replacement.

Treat a claim that a screening tool is objective as a claim to check, not a fact. Ask what evidence links the scores to how someone actually performs on the job, and whether the same rules apply to every applicant. The study itself cannot answer that. Its authors say the narrow cultural and institutional setting may limit how far the findings travel to other health care contexts, that all participants had recent experience with these systems, that interviews cannot objectively assess effectiveness, and that the findings may date quickly as the technology changes.

So read this as 20 people's perceptions in South Korea, not a scoreboard. No pass rate, no turnover figure, no evidence that the machine picks better or worse nurses appears here. The applicants and educators described a tool that was efficient and, in their view, not yet fair or transparent enough.

You may meet a version of this screen wherever you apply for work: an answer you never hear back about. The study gives you a question to carry into that room. Before you sit down for it, ask the employer, in writing, what the machine scores and whether anyone will tell you why you passed or failed.

Where each piece of context comes from, and how much of it we read

LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "these assessments (as specified by a leading developer in South Korea; Midas IT, 2023 ) generally consist of a self-introduction, simple video interview, personality test, situational questions, an online AI game (to evaluate reasoning and multitasking), and in-depth questions."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "A qualitative descriptive design was used in this study. Data were collected between February and August 2024 in South Korea. Semistructured focus group interviews were conducted with 10 nursing applicants and 10 nurse educators, including nursing professors and nurse managers, regarding their experience participating in AI-based competency assessments."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "They were frustrated because they did not know the exact assessment criteria used, the bases of the assessment outcomes, or whether the evaluation was made by a machine or a person."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "There are many rumors regarding the evaluation criteria. This was frustrating because both those who passed and those who failed did not know the reason . (Nursing applicant group 2-4)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "I think AI-based evaluations have a strong mechanical aspect that makes it difficult to reveal the warm human side of the students. (Professor group 1-2)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "I learned that nursing is a human-to-human interaction and a noble task. It is a contradiction to use AI to select nurses, who care for patients. (Nursing applicant group 1-1)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "I can prepare for face-to-face interviews by understanding applicant capabilities in advance. Thanks to AI, the burden of human resource evaluation work that I am responsible for has been reduced, and the fatigue caused by this has been alleviated. I think an environment has been created where I can focus more on my main job, nursing . (Manager group 1-2)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "AICAs and similar AI-based selection tools have been used in the United States and Japan since 2018, and are increasingly being adopted by institutions in South Korea ( Kim & Kwon, 2022 )."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "despite the use of AICAs in nurse selection in South Korea for 4 years, the effectiveness of AICAs in vetting and selecting nurses based on these competencies in the South Korean context has yet to be examined in the literature."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "When I was in my 4th year, if my classes and (nurse) job interviews overlapped, I had to give one of them up. However, I like that the AICA allows me to participate conveniently at any time and place I want, thus avoiding skipping class . (Nursing applicant group 1-3)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "Because so many people were connected at once, it kept crashing for three hours. I logged in early in the morning when there were not too many people, and I searched for a well-lit and soundproof internet café to avoid technical issues. I paid attention to all of the variables that could have an impact . (Nursing applicant group 1-5)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "There is bound to be a difference between students who can freely use information technology and those who cannot. I am concerned that an imbalance in information support may lead to fairness issues for applicants . (Professor group 2-3, Manager group 1-1)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "I heard that because the evaluation is done by a robot, you have to be as consistent as possible to obtain a high score. Instead of speaking in my own style, I practiced speaking in a stereotypical way, like a robot: “Hello, this is XXX.” (Nursing applicant group 1-1)"
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "To ensure validity, developing tailored algorithms that reflect core nursing competencies and establishing clear evaluation criteria is necessary. Also, interdisciplinary collaboration is needed to support the ethical and practical integration of AI in the process of selecting and hiring new nurses."
LEE, H., RYU, H., YOO, H. J. (2026). Is the Application of Artificial Intelligence-Based Competency Assessments for Selecting Clinical Nurses Acceptable? A Qualitative Study. Journal of Nursing Research. https://doi.org/10.1097/jnr.0000000000000768 - the article this story is about — the whole article — the passage: "I believe that the results of applicant evaluations using AI will provide useful basic data for face-to-face interviews . (Manager group 1-3)"

Do not take this as professional medical advice.

The findings of other studies mentioned here are known to us through this document, which is the one we read; we did not open each of those studies.

The free weekly drip by email

One email a week: the free story of every section, with its picture, and the headlines of the rest.

No payment, nothing to cancel. Unsubscribe from any issue.

El goteo semanal gratis por correo

Un correo por semana: la nota gratuita de cada sección, con su imagen, y los titulares de las demás.

Sin pago y sin nada que cancelar. Puede darse de baja desde cualquier envío.