Italy Asked Who Earns Money From Apps. The Poorest Were Not the Youngest.

survey · PloS one · la publicación, 30 jun 2026 · gratis

Quién termina haciendo mandados por aplicación, y quién apenas redondea el sueldo

Una encuesta oficial italiana de 2018 y 2021, leída con aprendizaje automático, encontró cuatro grupos distintos entre quienes trabajan por plataformas; no puede decir que el trabajo por aplicación empobrezca a nadie.

Versión breve · la versión detallada sigue, unos 9 min

Pregunte al chat de weeklyAI

Pregúnteme por este estudio: a quiénes se estudió, qué encontró y qué no dice.

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: Personas de 18 a 74 años encuestadas sobre trabajo por plataformas digitales
Cuántos: 785,480 trabajadores de plataforma estimados, dentro de una población de referencia de 86,687,651 personas en los dos años
Dónde: Italia
Cuándo: 2018 y 2021
Tipo de estudio: survey
Quién lo hizo: Scuola Normale Superiore de Pisa, Universidad de Pisa, Universidad de Bari y Universidad La Sapienza de Roma
El límite que importa: La encuesta solo capta a quienes trabajaban por plataformas en ese momento; quienes ya salieron no aparecen

Ser buscador de empleo y trabajar por plataforma, antes y después de la pandemia

En 2018frente aEn 2021

En 2018, ser buscador de empleo se asociaba con menor probabilidad de clasificar como trabajador de plataforma; en 2021, la misma condición pasó a ser un predictor positivo fuerte

La encuesta cuenta sólo a quienes trabajan por aplicación en el momento de responder.
Lectura de weeklyAI

Ver · 1:00

El Instituto Nacional de Análisis de Políticas Públicas de Italia, el INAPP, encuesta cada año a decenas de miles de personas de 18 a 74 años. En 2018 preguntó a 45,000; en 2021, a 46,282. Entre otras cosas, preguntó si la persona había ganado dinero aceptando encargos por sitios web o aplicaciones como Deliveroo o Glovo. Dijo que sí 222 personas en 2018 y 492 en 2021.

Con esos números, un equipo de investigadores italianos (Scuola Normale Superiore de Pisa, Universidad de Pisa, Universidad de Bari y Universidad La Sapienza de Roma) usó aprendizaje automático para ordenar a esas personas en grupos. Cuatro grupos aparecieron.

Dos de ellos apenas dependen de lo que ganan por aplicación: son jóvenes menores de 30 años que viven con sus familias, sin personas a su cargo. Para ellos la plataforma es un ingreso suelto. Los otros dos dependen de verdad de ese dinero: son mayores, con hijos y con parientes o conocidos a su cuidado, y muestran apuros de dinero, como postergar tratamientos médicos o dentales.

Entre los trabajadores con contratos no estándar, el modelo asoció el trabajo por aplicación con tener un segundo empleo, con contratos a plazo fijo o de tiempo parcial, con cuotas hipotecarias altas y con cuidar a parientes o conocidos que perdieron autonomía.

Entre quienes buscaban trabajo y tenían menos de 50 años, el trabajo por aplicación no apareció más entre los más pobres. Los ingresos familiares muy bajos y las cuotas hipotecarias altas apuntaban en sentido contrario. Y en 2021, estar buscando trabajo se volvió una señal fuerte de trabajar por aplicación; en 2018 no lo era.

La encuesta cuenta sólo a quienes trabajan por aplicación en el momento de responder. Quienes ya lo dejaron no aparecen, y los migrantes están muy poco representados: apenas el 3.8% de los encuestados no tiene ciudadanía italiana. Los casos más duros quedan fuera, así que la fragilidad real probablemente sea mayor que la que muestra este trabajo. Los modelos muestran asociaciones, no causas: no se puede decir que la aplicación haya empobrecido a nadie.

Vale la pena mirar si lo que entra por aplicación completa un sueldo que se encoge, o si lo reemplaza, y si sus horas y su contrato se están pareciendo a los de tiempo parcial o a plazo fijo.

Qué significa para usted

En Italia, la encuesta sugiere que quien depende de verdad de lo que gana por aplicación suele ser mayor, con hijos y con alguien a su cuidado, y no el joven que apenas redondea. Si usted vive de un oficio o de un empleo, fíjese si ese ingreso extra completa un sueldo que se encoge o si lo reemplaza, y pregunte cómo son sus horas y su contrato.

Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237

Quién pagó: Los autores fueron financiados por la Scuola Normale Superiore de Pisa, el proyecto PNRR FAIR, la Unión Europea mediante la subvención ERC 834756 (XAI) y el proyecto PRIN 2022 PNRR del Ministerio de Universidad e Investigación de Italia; el artículo no indica que los financiadores tuvieran algún papel en el estudio.

No tome esto como consejo médico profesional.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1703 palabras · unos 9 minLeerla →Cerrar

Trabajo por encargo en Italia: quién reparte, limpia y hace tareas en línea, según una encuesta analizada con inteligencia artificial

Un estudio con casi 800,000 trabajadores de plataforma estimados halla que el ingreso de plataforma suele completar un empleo inestable o perdido, no reemplazarlo. Los datos son italianos y no permiten saber qué pasó con quienes ya salieron.

Así podría verse · ilustración generada por weeklyAI.watch, no es una fotografía

En Italia, según una encuesta oficial analizada con aprendizaje automático e inteligencia artificial explicable, el trabajo en plataformas no es un grupo parejo sino varios. La vulnerabilidad económica no se concentra en los más jóvenes, como suele decirse, sino en personas mayores o más establecidas que enfrentan inestabilidad laboral, subempleo o ingresos que caen en sus ocupaciones tradicionales¹. El estudio lo firman investigadores de la Scuola Normale Superiore, la Universidad de Pisa, la Universidad de Bari y la Universidad de Roma "La Sapienza"; los datos los recogió INAPP, un instituto público italiano, sin participación de los autores, y el financiamiento vino de esas instituciones y de fondos públicos y europeos de investigación. Los autores declararon no tener conflictos de interés.

Conviene entender la escala de lo que se midió. En las dos encuestas —2018 y 2021, personas de 18 a 74 años— hubo 222 y 492 personas que dijeron haber ganado dinero por plataformas. Con los pesos estadísticos, eso equivale a unos 785,480 trabajadores de plataforma dentro de una población de referencia de 86.687.651 personas en los dos años². Es decir: una porción pequeña del total, y por eso los propios autores advierten que sus modelos trabajan con una minoría muy chica dentro de la muestra, lo que debilita la solidez de las predicciones.

¿Qué es una plataforma de trabajo digital aquí? La encuesta la define como "un sitio web o una aplicación móvil que conecta directamente con personas que solicitan sus servicios", y que exige crear un perfil de usuario para encontrar y aceptar tareas y cobrar cuando el servicio termina; entre los ejemplos citados están Deliveroo o Glovo³. El trabajo abarcó reparto de comidas y paquetes, tareas en línea como carga de datos o desarrollo de software, trabajo doméstico y traslados en auto.

El caso italiano tiene una particularidad que ayuda a leer los resultados: el artículo describe un estudio en el que el empleo no estándar creció allí más rápido que en la mayoría de los países europeos, con salarios estancados, más uso de contratos temporales y de tiempo parcial involuntario, y condiciones que se deterioran para los grupos vulnerables⁴. Las plataformas entraron sobre todo en sectores ya precarizados, como comercio, turismo y hotelería.

El hallazgo central es que el trabajo de plataforma funciona principalmente como un complemento, no como un oficio autónomo. Se asocia a contratos inestables, desigualdades de género e inseguridad económica, y opera como mecanismo compensatorio para completar ingresos insuficientes de empleos precarios⁵. En la práctica, quienes trabajan por plataformas suelen tener varias fuentes de ingreso inestables y, a menudo, un segundo empleo además de un trabajo de tiempo parcial o a plazo fijo⁶.

Aquí conviene ver el mecanismo, porque explica por qué esta forma de trabajar es distinta de un empleo común. El artículo describe un estudio en el que los trabajadores de plataforma suelen estar aislados, sujetos a monitoreo continuo, y con poca capacidad de entender o cuestionar las decisiones de los sistemas digitales que gestionan su trabajo⁷. El estudio recoge que los trabajadores de plataforma suelen estar aislados, sujetos a monitoreo continuo, y con poca capacidad de entender o cuestionar las decisiones de los sistemas digitales que gestionan su trabajo⁷. Y el artículo describe un estudio en el que enfrentan incertidumbre considerable sobre la estabilidad de sus ingresos y sobre la protección frente a riesgos sociales⁸.

El análisis por grupos encontró cuatro perfiles, no uno. Dos de ellos dependen mucho del ingreso de plataforma: uno con empleo precario, concentrado en el sur de Italia y las islas, con trabajos de tiempo parcial de pocas horas y baja paga pese a largas trayectorias; y otro de personas que buscan empleo, sobre todo en el norte. Ambos son en su mayoría hombres mayores, con responsabilidades de cuidado de hijos y de parientes con pérdida permanente de autonomía, y muestran señales de estrechez económica, como postergar tratamientos médicos. Los otros dos grupos dependen poco de la plataforma: personas jóvenes que viven con sus familias, para quienes es una actividad marginal o exploratoria.

Entre quienes buscaban empleo, la relación con las plataformas apareció más fuerte en personas jóvenes con presión económica moderada, no extrema; los ingresos familiares muy bajos y las cuotas hipotecarias altas aparecieron como predictores negativos⁹. También se asoció con haber postergado tratamientos médicos o dentales por falta de dinero¹⁰. Y los hombres y las personas sobrecalificadas para su puesto aparecieron algo más propensos a ser clasificados como trabajadores de plataforma, aunque el efecto es leve¹¹.

El año importa. En 2018, ser buscador de empleo se asociaba con menor probabilidad de clasificar como trabajador de plataforma; en 2021, la misma condición pasó a ser un predictor positivo fuerte¹². El estudio lee ese giro como señal de que la pandemia empujó a las plataformas a funcionar como opción de respaldo para desempleados, un amortiguador informal ante menos oportunidades y una inseguridad de ingresos persistente¹³.

Vale la pena comparar dos estudios que leímos solo en su resumen, porque el texto completo está detrás de una suscripción. Uno, hecho en Wuhan, China, sobre el servicio de robotaxis de Baidu, halló que la llegada de esos vehículos redujo el ingreso diario promedio de los taxistas tradicionales un 10.9% y, según una encuesta a esos conductores, aumentó sus horas de trabajo y su estrés laboral¹⁴¹⁵¹⁶. El otro, con datos de la encuesta de fuerza laboral de Canadá, encontró que las ocupaciones más expuestas a los grandes modelos de lenguaje son justamente aquellas donde la precariedad es más baja, ocupaciones que antes habían quedado al margen del cambio tecnológico¹⁷¹⁸¹⁹.

Así lo leemos nosotros. Cuando el trabajo se organiza por encargos sueltos en vez de un puesto estable, la gente encadena actividades distintas, con huecos sin ingreso entre una y otra, y deja de pensar su vida laboral como una carrera que avanza. Esperaríamos, entonces, que entre quienes sacan dinero de plataformas los ingresos lleguen en tramos irregulares, alternados con períodos sin nada, incluso en personas con estudios y varios oficios, y que muchas describan la plataforma como un relleno entre trabajos y no como su oficio principal. Sabríamos que nos equivocamos si la mayoría de esas personas reportara ingresos parejos mes a mes, un solo trabajo, y la plataforma como su ocupación central y duradera. Lo que usted puede hacer con esto: si depende de encargos sueltos, lleve su propio registro de meses con y sin ingreso y calcule cuánto necesita guardar en los meses buenos para cubrir los huecos. Y al pedir trabajo, pregunte cuántos encargos o contratos cortos han encadenado antes de ofrecerle este.

Y hay una segunda cosa que nos parece igual de importante. Cuando las reglas de pago, calidad y reclamo las aplica un sistema automático y no una persona, quien trabaja queda sin interlocutor al que reclamar, y esa falta de respaldo pesa más sobre quienes recién entran al empleo, sobre quienes cargan el cuidado de la familia y sobre quienes no pueden firmar un contrato formal. Esperaríamos que quienes más dependen del ingreso de plataforma sean personas con responsabilidades de cuidado y con historial de contratos inestables, y que sus reclamos se resuelvan por reglas automáticas sin alguien que responda. Nos equivocaríamos si esas personas resolvieran sus reclamos con un supervisor humano identificable, o si la dependencia del ingreso no se concentrara en quienes cuidan a otros. Antes de aceptar encargos, averigüe por dónde se reclama un pago mal hecho y si responde una persona. Guarde capturas de pantalla de cada tarea, tarifa y mensaje: es su único respaldo si el sistema decide solo.

Nos queda una advertencia sobre lo que estos datos no pueden decir, y es grande. La encuesta solo capta a quienes trabajaban por plataformas en ese momento: quienes ya salieron no aparecen, así que el cuadro puede verse mejor que la historia completa de este trabajo²⁰. Además, los migrantes y los trabajadores informales están muy mal contados —solo el 3.8% de los encuestados no tiene ciudadanía italiana, y de ellos apenas el 1.6% viene de fuera de la Unión Europea—, de modo que los casos más duros probablemente faltan y la fragilidad real quedaría subestimada²¹. Y los modelos muestran correlaciones, no causas: el estudio no puede decir que el trabajo de plataforma empobreció a alguien, ni que dejarlo lo ayudaría.

Los propios autores señalan otros límites: la proporción de trabajadores de plataforma en las submuestras analizadas va del 1.5% al 4.5%, una minoría muy chica que debilita los modelos; no hay seguimiento en el tiempo de las mismas personas; y el análisis no cubre las diferencias de condiciones entre plataformas ni los factores del lado de quien contrata. También aclaran que sus resultados valen para Italia y no se trasladan sin más a otros mercados laborales.

¿Qué haría falta después? Los autores piden encuestas propias con diseño de panel, es decir, seguir a las mismas personas en el tiempo, y un registro administrativo más sistemático del trabajo mediado por plataformas, porque las tareas en línea y el trabajo doméstico y de cuidado pueden quedar fuera de los recuentos. También proponen reforzar la protección social para desempleados y para quienes tienen empleo no estándar, y políticas de conciliación, dado que muchos trabajadores de plataforma cargan tareas de cuidado.

¿Y qué significa esto donde usted vive? El estudio es italiano y no dice nada sobre su país. Pero la pregunta que deja es transportable: cuando alguien de su casa empiece a repartir, limpiar o hacer tareas en línea, la primera pregunta útil no es cuánto paga por hora, sino si eso viene a reemplazar un ingreso que se cayó o a completar uno que no alcanza. Esa diferencia cambia por completo lo que hay que reclamar: un empleo, o protección para un ingreso que ya existe pero no llega a fin de mes. Mírelo así la próxima vez que alguien de su familia firme con una aplicación: ¿este trabajo es el piso, o es el parche?

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

Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Economic vulnerability is concentrated not among the youngest workers, as often suggested in the literature, but among older or more established individuals facing job instability, underemployment, or declining income from traditional occupations."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Within these samples, 222 and 492 respondents, respectively, reported engaging in digital platform work. After applying survey weights, these figures correspond to an estimated 785,480 platform workers out of a total reference population of 86,687,651 individuals across the two years."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The PLUS survey defines a digital labour platform as a “a website or mobile app that connects directly with people who request their services. These platforms (e.g., Deliveroo or Glovo) require workers to create a user profile to find and accept tasks or assignments and to receive payment once the service has been completed”"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "In Italy, non-standard work has expanded more rapidly than in most European countries, accompanied by stagnant wages, increased reliance on temporary and involuntary part-time work, and deteriorating conditions for vulnerable groups"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "platform work is associated with structural vulnerabilities typical of non-standard employment, including unstable contracts, gender inequalities, and economic insecurity, and it primarily functions as a compensatory mechanism to supplement insufficient earnings from precarious jobs."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Platform workers are typically associated with multiple and unstable income sources, often holding a secondary job in addition to part-time or fixed-term employment."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "platform workers are frequently isolated, subject to continuous monitoring, and have limited capacity to understand or challenge decisions made by the digital systems that manage their work"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Platform workers often experience precarious employment relationships and face considerable uncertainty regarding income stability and protection against social risks"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Among jobseekers, engagement with platforms is more likely among younger individuals experiencing moderate—rather than severe—financial strain, suggesting that platform work is not generally perceived as a last-resort strategy but rather as a temporary or adaptive response to limited labour market opportunities."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Jobseekers who had to postpone medical treatments (including dental care) for financial reasons, or who report a limited ability to cope with sudden expenses, exhibit a positive association with platform work."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "men and overqualified workers are slightly more likely to be classified as platform workers by the model, though these effects are less pronounced."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "in 2018, being a jobseeker was associated with a lower probability of being classified as a platform worker; by contrast, in 2021, the same condition became a strong positive predictor."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The COVID-19 pandemic further intensified these dynamics, acting as a catalyst for workers experiencing economic and social stress. During this period, platform work expanded as a fallback option for the unemployed, providing an informal buffer amid declining employment opportunities and persistent income insecurity."
Yu Z, Wang J, Zuo T, Alm J, Li X, Li X. (2026). Robotaxis reduce taxi drivers’ income. Humanities and Social Sciences Communications. 10.1057/s41599-026-07366-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "This study investigates the labor market impact in Wuhan, China, of Baidu’s Apollo Go, a robotaxi service."
Yu Z, Wang J, Zuo T, Alm J, Li X, Li X. (2026). Robotaxis reduce taxi drivers’ income. Humanities and Social Sciences Communications. 10.1057/s41599-026-07366-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "We find that the introduction of robotaxis reduces traditional taxi drivers’ average daily income by 10.9%, likely due to the reduced demand for their services."
Yu Z, Wang J, Zuo T, Alm J, Li X, Li X. (2026). Robotaxis reduce taxi drivers’ income. Humanities and Social Sciences Communications. 10.1057/s41599-026-07366-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "In addition, we conduct a survey of these traditional taxi drivers, and we find that robotaxis increase working hours, increase job stress, decrease job satisfaction, and encourage these traditional taxi drivers to seek alternative employment."
Jetha A, Liao Q, Smith P, Vu V, Biswas A, Smith B, et al. (2026). Large language model exposure and precarious occupations: Unpacking relationships in the Canadian labor force. Scandinavian Journal of Work, Environment & Health. 10.5271/sjweh.4312 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Large language models (LLM) may reshape this relationship. This study examines the association between occupational exposure to LLM and occupational precarity."
Jetha A, Liao Q, Smith P, Vu V, Biswas A, Smith B, et al. (2026). Large language model exposure and precarious occupations: Unpacking relationships in the Canadian labor force. Scandinavian Journal of Work, Environment & Health. 10.5271/sjweh.4312 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Using Canada's Labour Force Survey, occupational exposure to LLM and four dimensions of precarity (contractual instability, earnings inadequacy, schedule unpredictability, working-time mismatch) were examined."
Jetha A, Liao Q, Smith P, Vu V, Biswas A, Smith B, et al. (2026). Large language model exposure and precarious occupations: Unpacking relationships in the Canadian labor force. Scandinavian Journal of Work, Environment & Health. 10.5271/sjweh.4312 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Occupations most likely to be exposed to LLM are those where precariousness is lowest. These occupations have previously been sheltered from technological change."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "our cross-sectional design introduces survivorship bias: the PLUS survey captures only current platform workers, while those who have already exited remain unobserved."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - el artículo del que trata esta nota — el artículo completo — el pasaje: "only 3.8% of respondents do not possess Italian citizenship, and of these, just 1.6% come from non-European Union countries. This limitation implies that the specific dynamics of platform work among migrants are insufficiently represented"

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.

survey · PloS one · the paper, 30 Jun 2026 · free

Italy Asked Who Earns Money From Apps. The Poorest Were Not the Youngest.

Two survey years, about 785,000 platform workers, and a warning: the survey only sees people still doing the work.

Short version · the longer version follows, about 7 min

Ask the weeklyAI chat

Ask me about this study: who was studied, what it found, and what it does not say.

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: People in Italy who earned money through websites or apps
How many: 785,480 platform workers out of a total reference population of 86,687,651 individuals across the two years
Where: Italy
When: 2018 and 2021
Kind of study: survey
Who did it: Scuola Normale Superiore in Pisa and three Italian universities
The limit that matters: It counts only people doing platform work now, so those who already quit are invisible

Being a jobseeker and being classified as a platform worker, 2018 against 2021

In 2018againstIn 2021

In 2018, being a jobseeker was linked to a lower chance of being classified as a platform worker; in 2021, that flipped, and being a jobseeker became a strong positive predictor

In Italy, the strain sits with older workers whose regular jobs shrank, not with the youngest.
weeklyAI's reading

Watch · 0:57

A public research body in Italy, the National Institute for Public Policy Analysis, runs a survey called PLUS. In 2018 it asked 45,000 people whether they earned money through websites or apps like Deliveroo or Glovo. In 2021 it asked 46,282 people the same question. Of those, 222 and 492 said yes. The survey counts them, with weights, as about 785,480 platform workers across both years.

Researchers from three Italian universities and the Scuola Normale Superiore in Pisa then let machine-learning programs sort those workers into groups. Four came out. Two groups barely depend on the app money. They are younger, under 30, mostly still living with their parents, and the work is a small extra for them. Two groups depend on it heavily. Those workers are older, often supporting children and caring for relatives. Many have put off medical or dental treatment for lack of money.

That result runs against the usual story. In Italy, the strain sits with older workers whose regular jobs shrank, not with the youngest.

Among people in part-time, fixed-term or otherwise non-standard jobs, platform workers often had a second job, high mortgage costs and skipped care. The study cannot say the apps caused any of it. It only shows the two things appear together.

Among jobseekers under 50, platform work was more common for those under moderate money pressure. The very poorest were less likely to be doing it.

Two things the survey misses matter. It only counts people doing platform work now, so those who already quit are invisible, and the picture may look gentler than the full history. Migrants are barely counted: 3.8 percent of respondents were not Italian citizens. The hardest cases are likely missing.

Watch your own trade. If app or gig earnings are topping up a paycheck that keeps shrinking, or replacing it, that is the pattern this survey found.

What this means for you

The study is about Italy, so treat it as a hint about your own trade rather than a rule for it. If app or gig earnings are topping up a paycheck that keeps shrinking, or standing in for one, that is the pattern this survey found, and it says nothing about which one came first.

Who paid: The authors were funded by Scuola Normale Superiore of Pisa, the PNRR FAIR project, the European Union under ERC grant 834756 (XAI), and the Italian Ministry of University and Research PRIN 2022 PNRR project; the article does not say the funders had any role in the study.

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 longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1381 words · about 7 minRead it →Close

App Work in Italy Tops Up a Shrinking Paycheck, Not a New Career

A national survey finds the people leaning hardest on delivery and task apps are older workers whose old jobs were fading.

How it could look · illustration generated by weeklyAI.watch, not a photograph

In Italy, the people who depend most on app-based work are not the youngest. They are older or more established workers whose regular jobs became unstable, who were working fewer hours than they wanted, or whose income from ordinary work was falling¹. A team led by Clara Punzi of the Scuola Normale Superiore in Pisa reached that conclusion by analysing two waves of a national labour survey, INAPP PLUS, from 2018 and 2021, and feeding the answers through machine learning and explainable AI tools.

The survey interviewed 45,000 people in 2018 and 46,282 in 2021, all aged 18 to 74. Of those, 222 and 492 people respectively said they earned money through a website or mobile app. Weighted to the whole population, that represents an estimated 785,480 platform workers out of a total reference population of 86,687,651 across the two years². The survey reports the activities platform workers do: delivering products or packages, delivering meals, performing online tasks like data entry and software development, domestic work, driving someone by car, and other unspecified activities. The numbers are small enough that the authors could not separate those activities in the analysis.

The study's second finding is about what platform work sits next to. It lines up with unstable contracts, gender inequality and income insecurity, and it mostly works as a way to top up earnings that are not enough on their own³. Platform workers typically hold several income sources at once, often a second job alongside part-time or fixed-term work⁴. Men and workers who say their skills are above what their job requires were slightly more likely to be classified as platform workers⁵.

Among people looking for work, the pattern is narrower than you might expect. Those turning to platforms were more often younger and under moderate financial strain rather than severe strain, which the authors read as a temporary or adaptive response to thin job options, not a last resort⁶. Jobseekers who had put off medical or dental treatment for money reasons, or who said they could not handle a sudden expense, were more likely to be classified as platform workers⁷.

The pandemic shows up as a turning point. In 2018, being a jobseeker was linked to a lower chance of being classified as a platform worker; in 2021, that flipped, and being a jobseeker became a strong positive predictor⁸. The authors describe the pandemic as a catalyst for workers already under economic and social stress, with platform work expanding as a fallback for the unemployed and acting as an informal buffer as opportunities shrank⁹.

It helps to be clear about what a "digital labour platform" is in this research. The survey defines it as a website or mobile app that connects directly with people who want a service, where the worker creates a profile to find and accept tasks and gets paid once the job is done; the survey names Deliveroo and Glovo as examples¹⁰. That is the whole mechanism: the app is the middleman, the profile is the door, and the payment arrives after the task closes.

Italy makes an unusually clear test case. Non-standard work there grew faster than in most European countries, alongside wages that did not move, more temporary and involuntary part-time work, and worse conditions for people already exposed¹¹. The country is not a stand-in for anywhere else, and the study does not claim to be.

The authors also lay out what is known about this kind of work from the wider research they review. Platform workers often face uncertain income and weak protection against social risks¹². They are frequently isolated, watched continuously, and have little ability to understand or challenge the decisions the digital systems make about their work¹³. The study itself did not measure those conditions; it cites them.

The limits matter as much as the findings. The survey counts only people doing platform work at the moment, so those who already quit are invisible, and the picture may look better than the full history¹⁴. Migrants are badly under-counted: only 3.8% of respondents are not Italian citizens, and of those just 1.6% come from outside the European Union, so the hardest cases are likely missing and the real fragility understated¹⁵. The models show associations, not causes; the authors state plainly that their explainability method cannot establish causal relationships and that its results must be read as correlational. Platform workers were a very small minority of the subsamples used for the two classification tasks, and the authors say this imbalance may have weakened the models. And the analysis says nothing about differences in working conditions between one platform and another.

Here is how we read it. When a household loses stable ground, it tends to take the arrangement that lowers the worst outcome, even when that arrangement gives away most of the upside and leaves the household dependent on whoever controls the work. That is what this survey's shape suggests to us. If we are right, the people leaning hardest on app work in a place like yours are those who already lost something — a shift, a contract, a wage that used to cover the month — and they accept thinner terms because the alternative is nothing. You would know we are wrong if the heaviest users turned out to be people with steady jobs and savings, adding a little on the side, or if they reported better pay and security than other workers in unstable jobs. If you or someone close to you starts app work after a job loss, treat it as a stopgap and keep a written record of hours, pay and expenses, so you can see whether it is covering the gap or quietly becoming the main income.

We also read this as a story about who holds the record. When contracts and official channels cannot be relied on, people fall back on relatives, neighbours and others from the same place, and they spend unpaid time checking ratings, payments and account status instead of building anything. If that holds, platform workers in Italy — migrants and those in the south especially — would find work through their own networks and spend real hours watching the app rather than being watched over by a boss. You would know we are wrong if they mostly found work through job centres or open hiring and reported no extra unpaid time spent chasing payment or account problems. Before you sign up to a platform, ask around your own network, and keep screenshots of accepted jobs, pay and messages; in informal work the record you keep yourself is often the only proof you have.

One more thing is worth watching. People tend to go along with whoever is presented as the one in charge, even when the arrangement costs them, and they rarely question the arrangement itself. If that holds here, platform workers in Italy would seldom challenge what the app decides about pay, ratings or a suspended account, and would treat it as normal. You would know we are wrong if they appealed regularly, organised together, or switched platforms when treated unfairly. Next time an app or a dispatcher gives you an instruction that seems wrong, write down what was asked and what happened, and ask a co-worker whether they have seen the same thing; a shared record is harder to ignore than a single complaint.

For anyone outside Italy, the useful move is to ask the same questions locally rather than assume the answers travel. The finding that older, more established workers carry the strain is a fact about Italy in these two survey years, and the study is explicit that it is focused on that labour market.

What the study makes possible is a different question to bring to your own situation. When app work enters a household after a job loss, the thing to watch is not whether the app pays something — it usually does. It is whether the hours, the expenses and the waiting are quietly turning a stopgap into the main income, and who then holds the record of what you were owed.

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

Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Economic vulnerability is concentrated not among the youngest workers, as often suggested in the literature, but among older or more established individuals facing job instability, underemployment, or declining income from traditional occupations."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Within these samples, 222 and 492 respondents, respectively, reported engaging in digital platform work. After applying survey weights, these figures correspond to an estimated 785,480 platform workers out of a total reference population of 86,687,651 individuals across the two years."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "platform work is associated with structural vulnerabilities typical of non-standard employment, including unstable contracts, gender inequalities, and economic insecurity, and it primarily functions as a compensatory mechanism to supplement insufficient earnings from precarious jobs."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Platform workers are typically associated with multiple and unstable income sources, often holding a secondary job in addition to part-time or fixed-term employment."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "men and overqualified workers are slightly more likely to be classified as platform workers by the model, though these effects are less pronounced."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Among jobseekers, engagement with platforms is more likely among younger individuals experiencing moderate—rather than severe—financial strain, suggesting that platform work is not generally perceived as a last-resort strategy but rather as a temporary or adaptive response to limited labour market opportunities."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Jobseekers who had to postpone medical treatments (including dental care) for financial reasons, or who report a limited ability to cope with sudden expenses, exhibit a positive association with platform work."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "in 2018, being a jobseeker was associated with a lower probability of being classified as a platform worker; by contrast, in 2021, the same condition became a strong positive predictor."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "The COVID-19 pandemic further intensified these dynamics, acting as a catalyst for workers experiencing economic and social stress. During this period, platform work expanded as a fallback option for the unemployed, providing an informal buffer amid declining employment opportunities and persistent income insecurity."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "The PLUS survey defines a digital labour platform as a “a website or mobile app that connects directly with people who request their services. These platforms (e.g., Deliveroo or Glovo) require workers to create a user profile to find and accept tasks or assignments and to receive payment once the service has been completed”"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "In Italy, non-standard work has expanded more rapidly than in most European countries, accompanied by stagnant wages, increased reliance on temporary and involuntary part-time work, and deteriorating conditions for vulnerable groups"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "Platform workers often experience precarious employment relationships and face considerable uncertainty regarding income stability and protection against social risks"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "platform workers are frequently isolated, subject to continuous monitoring, and have limited capacity to understand or challenge decisions made by the digital systems that manage their work"
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "our cross-sectional design introduces survivorship bias: the PLUS survey captures only current platform workers, while those who have already exited remain unobserved."
Punzi, C., Cirillo, V., Guarascio, D. et al. (2026). Platform workers not by chance: Exploring the digital labour markets in Italy with machine learning and explainable AI. PLOS One. https://doi.org/10.1371/journal.pone.0340237 - the article this story is about — the whole article — the passage: "only 3.8% of respondents do not possess Italian citizenship, and of these, just 1.6% come from non-European Union countries. This limitation implies that the specific dynamics of platform work among migrants are insufficiently represented"

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.