Revista de economía mundial 69, 2025, 95-116
ISSN: 1576-0162
DOI: http://dx.doi.org/10.33776/rem.v0i69.8289
Job AutomAtion in brAzil: Which consequences for Women’s
employment?
AutomAtizAción del empleo en brAsil: ¿qué consecuenciAs pArA
el empleo femenino?
Kethelyn Ferreira
kethelynff@gmail.com
Universidade Federal do Rio de Janeiro
Marta Castilho
castilho@ie.ufrj.br
Universidade Federal do Rio de Janeiro
Recibido: mayo 2024; aceptado: febrero 2025
AbstrAct
The automation of tasks and/or activities that constitute a work process
has the potential to replace human labor, raising concerns about technological
unemployment. Furthermore, individuals and groups who develop the
technology are not gender neutral. This paper explores the relationship
between advancements in job automation and gender inequalities within the
Brazilian paid labor market. In aggregate terms, the probability of automation
is lower for women than for men in Brazil. However, occupations with lower
probability of automation generally tend to be lower-skilled and associated
with lower pay – this is particularly true for most women's occupations, which
are linked to the care economy and reproductive tasks.
Keywords: Gender inequalities, job automation, Brazil, labor market.
resumen
La automatización de tareas y/o actividades laborales tiene el potencial
de reemplazar el trabajo humano, generando así preocupaciones acerca del
desempleo tecnológico. El presente artículo busca explorar la relación entre los
avances en la automatización laboral y las desigualdades de género en el mercado
laboral remunerado en Brasil. A nivel agregado, la probabilidad de automatización
para las mujeres es menor que para los hombres en Brasil. No obstante, las
ocupaciones con menor probabilidad de automatización suelen requerir menos
cualificación y están asociadas con salarios más bajos – allí se encuentran la
mayoría de las ocupaciones femeninas vinculadas a la economía del cuidado y a
las tareas reproductivas.
Palabras clave: Desigualdades de género, automatización del trabajo,
Brasil, mercado laboral.
JEL Classification/ Clasificación JEL: B54; E24; O33.
Revista de economía mundial 69, 2025, 95-116
1. introduction
The automation of tasks and/or activities that make up a work process
(hereinafter also referred to as “automation of jobs or occupations”) has
the potential to replace human labor, which could lead to concerns about
increased technological unemployment or change the characteristics of jobs
and the skills necessary for their performance. Furthermore, labor relations can
also be significantly impacted by changes in production processes, requiring
a new work organization in salaried employment and emerging types of work
(Mokyr et al., 2015).
Some authors affirm that automation processes can translate into
opportunities and create expectations of improvements in the quality of life
of workers (Mokyr et al., 2015). However, they also imply risks concerning
a deepening of current inequalities or the emergence of new inequalities.
After all, certain social groups (such as women and/or people from lower-
income brackets) could be less prepared to take advantage of such potential
opportunities (CEPAL, 2019a).
The people and groups who develop automation technology and look for
ways to take advantage of it are not gender-neutral (Roberts et al., 2019).
Consequently, the development or improvement of automation technologies is
also not neutral to gender inequalities in economies or how people integrate
into society.
How the automation of tasks in the labor market transforms the economy
and who benefits from it depend on several factors, such as the sectors where
automation occurs; what skills will be demanded in the future, and how they
are valued; who can adapt to new roles; and how gains are distributed (Roberts
et al., 2019).
Gender gaps threaten equal participation of women and men in the new
work paradigm emerging from technological progress and the introduction of
automation at work, to the detriment of female participation (CEPAL, 2019a).
Gender gaps also appear in the development of digital skills (Vaca Trigo and
Valenzuela, 2022), which are crucial for taking advantage of opportunities that
may arise from technological advancements in the labor market. Besides that,
people can also differ either by having different colors or races or dedicating
more or fewer hours to unpaid work and care, among other factors (Fontana,
98 Kethelyn Ferreira · Marta Castilho
2003), which reinforces the gender hierarchy perpetuated by the sexual
division of labor.1
This paper examines the relationship between advances in job
automation and gender inequalities in Brazil’s labor market. It is organized
into four sections, including the introduction and final considerations. First,
we theoretically explore gender inequalities in the labor market under the
automation era. Secondly, we discuss job automation and its potential impact
on workers. Thirdly, we outline the methodology for assessing the probability
of automation in various occupations. Finally, we examine which occupations
are most vulnerable to automation through a gender lens, using 2022 Brazilian
labor market data.
2. Gender inequAlities in the lAbor mArket durinG the AutomAtion erA
Gender inequalities and the naturalization of the subordinate role of women
to men present in societies result from a social construction passed down
through generations (Saffioti, 1987). The patriarchal system, which precedes
but intensifies during capitalism, is responsible for the formation of gender
roles (Lerner, 2019).
Patriarchy often assigns the role of caregivers to women, reinforcing social
expectations that direct them towards responsibilities centered on caring for
the family and home (Lerner, 2019). While most men direct their time and
effort to paid work, most women combine paid and unpaid work. Despite
their importance, social reproduction tasks are commonly unremunerated,
which contributes to reaffirming their undervaluation in our society (Melo and
Castilho, 2009).
The devaluation of women also manifests through disparities in the paid
job market. As Elson (1999) points out, the labor market is a “gender-bearing
institution”, responsible, in many aspects, for reinforcing discrimination against
women. Gender stereotypes permeate job market relations and structure,
where different types of work are categorized as “men’s jobs” and “women’s
jobs”.
The weight of unpaid domestic responsibilities placed on women penalizes
them within the labor market and is one of the main determinants of this
group occupying an underprivileged position in terms of both salary gains
and occupations. For example, given the socially conferred responsibility
of carrying out unpaid work, women tend to concentrate on the seasonal
workforce and consequently face fewer opportunities to update their human
capital or advance in their work (Barrientos, 2001).
1 In summary, oppression varies among women and is shaped by intersecting factors like gender, class,
and race (Davis, 2016). In Brazil, due to its history of slavery, racism and sexism are deeply intertwined,
disproportionately affecting Black women. Historically, Black women have been responsible for both
reproducing the labor force and performing highly exploited labor. Distinct gendered roles have been
created for Black and white women, with Black women facing objectification, sexualization, and the
denial of their political agency (Gonzalez, 1980).
99
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
Female segregation is a crucial factor of gender inequality in the labor
market, highlighting the limitations that women face when seeking employment
opportunities. Occupational segregation is divided into horizontal segregation,
associated with feminized and masculinized sectors, and vertical segregation,
when women are excluded from positions or occupations associated with
better wages or decision-making.
Compared to men, women typically face wage gaps and fewer opportunities
for advancement. It is worth noting that, for women, wage gaps are often
associated with the low value attributed to women’s work, reflecting the
historical sexual division. This disparity is a result of discrimination rather than
productivity differentials (Teixeira, 2016).
Technological changes and advancements are not inherently neutral
to gender inequalities. On the one hand, technological changes reflect and
reinforce preexisting distinctions and socio-economic structures in our society.
On the other hand, gender inequalities influence the ways women can take
advantage of digital innovations, placing them at a disadvantage in a scenario
of increasing technological advances in the labor market. Thus, we can consider
a bidirectional relationship between gender inequalities and technological
changes.
In the first direction, as pointed out by Howcroft and Rubery (2019), when
automated eligibility systems and predictive analytics are built using data
that contains gender biases, they can incorporate, perpetuate, and amplify
these biases within software design and operation. The authors refer to this
phenomenon as “bias in, bias out” and illustrate it with two examples: (i)
women being less likely to be shown ads for high-paid jobs on Google, and (ii)
Amazon’s recruitment AI penalizing resumes that included the word “women.”
In the second direction, the overload of unpaid care tasks affects women’s
digital skills development, as they have less time than men to explore
cyberspace. This disparity translates into higher rates of digital illiteracy
among women, meaning they have fewer skills to understand, control, and
establish trusting relationships with technology (Vaca Trigo and Valenzuela,
2022). In several countries in Latin America, women make more limited use
of digital technologies and engage in activities that require less technological
skills, which puts them at a disadvantage compared to men. The gender gap in
digital skills is significant, with women having 1.6 times less chance than men
of possessing such skills (UNESCO, 2019).
The low participation of women as internet users and in the fields of
information and communication technology (ICT) is closely related to a
patriarchal culture that discourages the development of digital skills. Gender
stereotypes manifest in norms, family pressures, and the lack of role models,
leading to the perception of technology as a predominantly male field. This
perception contributes to girls’ insecurity regarding their own digital skills
from an early age, affecting the inclusion of women and girls in science,
technology, engineering, and mathematics (STEM). Furthermore, the lack of
100 Kethelyn Ferreira · Marta Castilho
female representations in the tech sector, in educational materials, media and
advertising reinforces this perception (CEPAL, 2019b).
3. Job AutomAtion And its possible impActs on Workers
Technological innovations occur exponentially and cause transversal
transformations in economies and societies, impacting complete production,
management, and governance systems (CEPAL, 2018). Innovations, including
automation technologies, can translate into opportunities for economies and
imply risks regarding the increase or emergence of inequalities (CEPAL, 2019b).
The continuous advancement of artificial intelligence and robotics has
allowed machines to perform and automate more and more tasks, whether
routine, non-routine, physical, or cognitive.2 Activities previously considered
exclusive to humans can be performed more efficiently and economically
by machines (CEPAL, 2019b; McKinsey Global Institute, 2019; Roberts et
al., 2019). Although automation is not a new phenomenon - sectors such
as agriculture and manufacturing have experienced major replacements of
labor by machines in the past - the computerization of white-collar services
(associated, for example, with bureaucratic or management activities) in
advanced economies has accelerated (Acemoglu and Restrepo, 2018; Frey
and Osborne, 2016).
In recent decades, the replacement of several jobs by computers has
been evident, including functions as supermarket cashiers, clothing retail
store cashiers, and telephone operators. In general, labor-saving automation
is already eliminating many jobs that involve routine tasks requiring low and
medium skills, and task automation is expected to have an even broader
scope in the medium term (Brussevich et al., 2019). Progress in machine
learning has further expanded the set of activities that can be performed more
efficiently by computers than by humans (Brynjolfsson et al., 2018). Currently,
automation already goes beyond routine tasks associated with manufacturing.
One example is driverless autonomous cars, which indicate how manual tasks
in transportation and logistics could soon be automated (Brynjolfsson and
McAfee, 2011).
As machine-learning techniques progress, such as advances in digitalization
and artificial intelligence, the number of automatable tasks in the workplace is
increasing. These technological innovations could change how the production
2 Tasks can be classified into non-routine analytical, non-routine interactive, routine cognitive, routine
manual, and non-routine manual. Some examples of non-routine analytical tasks include research
and planning, while non-routine interactions involve negotiation and management. Routine cognitive
tasks cover activities like accounting, while routine manual tasks involve operating machines. Non-
routine manual tasks include repairing machines or restoring art. In this sense, routine manual tasks
are more susceptible to automation than non-routine analytical or interactive tasks (Black and Spitz-
Oener, 2010).
101
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
process is carried out, leading to increased productivity and growth. However,
they can also transform the nature of work itself (Brussevich et al., 2018).
Another way the automation of activities can affect the job market is
through the precariousness of working conditions. In situations where there
is no evidence of job losses, wages may decrease or stagnate due to a loss
of employee bargaining power. In developing economies, for example, the
relocation of industrial production can pressure wages (CEPAL, 2019b). On the
other hand, as Boddy et al. (2015) found, the automation of tasks in developed
countries has caused the average wages of workers to stagnate while the wages
of high-skilled workers continued to grow.
Precariousness can also manifest in more flexible work arrangements,
with weaker links between employers and workers, often restricting access to
traditional social protection mechanisms. Furthermore, varied atypical forms of
work many also emerge or be adopted, with new determinations of spaces and
working hours, such as intermittent work and variable hours (Novick, 2018).
The likelihood of automating work activities and thus eliminating jobs or
changing labor relations depends on the type of occupation performed by
workers and the associated responsibilities (Brussevich et al., 2018). The risk
of job destruction depends on the technological feasibility of replacing human
labor, but other economic, political, and social factors will also shape the
future of work. In short, economic, political, and social actors will decide the
future of work, but their actions are conditioned by the characteristics of new
technologies and their competitive use (Weller et al., 2019).
Furthermore, the impacts on jobs and labor relations are uncertain and
may vary across countries and population groups (CEPAL, 2018). According
to McKinsey Global Institute (2019), emerging economies are expected to
experience lower levels of automation relative to the size of their employed
population than mature economies, given the feasibility of technological
implementation. Developing countries usually face a range of barriers to
the absorption of innovations, which can result in a slower adoption of new
technologies (CEPAL, 2019b). The introduction of technological changes in
Latin America and the Caribbean, for example, is subject to a series of barriers.
The region lacks the necessary supply of skills and capabilities to meet the
demand of the ongoing technological revolution, thus preventing the adoption
of new technologies. Furthermore, the presence of smaller companies and low
wages in some occupations in the region limit innovation due to the high costs
of technology (CEPAL, 2019b).
The literature also suggests that, while jobs can be eliminated due to
advances in automation, demand is created for workers in non-automated
tasks due to productivity gains (Acemoglu and Restrepo, 2018; CEPAL,
2019a). In practice, the automation of activities would have two competing
effects on employment. First, as technology replaces labor, there is a knock-
on effect, requiring male and female workers to reallocate their labor supply.
Secondly, there is the capitalization effect, as more companies enter industries
where productivity is relatively high, causing employment in these industries to
102 Kethelyn Ferreira · Marta Castilho
expand (Frey and Osborne, 2016). In short, jobs maintained and/or created are
expected to be those most intensive in technical skills, cognitive skills, creative
activities, decision-making, managerial, and care functions (Brussevich et al.,
2018; Roberts et al., 2019).
Nevertheless, the potential positive impacts of automation in the labor
market on the economy and specific groups depend on several factors: who
can access the new jobs that may arise, what will happen to wages, what
changes to job conditions will occur or remain, and how “abundance”3 created
by increased productivity will be distributed (Roberts et al., 2019). In short,
the balance between job conservation and technological progress reflects
the balance between the distribution of power in society and the gains from
technological progress (Frey and Osborne, 2016).
As automation reconfigures the labor market, one relevant phenomenon
to observe concerns the “polarization of the labor market.” Since it affects
mostly routine tasks in “intermediary” occupations regarding remuneration
and qualification, the advance of automation can induce a concentration of
jobs in the extremes of employment based on qualification. This phenomenon
has been identified in many countries, from developed to developing ones.4
Rocha and Vaz (2023) found that the technologies adopted by manufacturing
industries in Brazil replace low-skilled workers in performing routine tasks while
complement workers with higher qualifications. Finally, other hypotheses help
explain such phenomenon, such as the Skills-Biased Technological Change
(SBTC) and the Routine-Biased Technological Change (RBTC) ones. The SBTC, as
discussed by Acemoglu and Autor (2011) and Goos et al. (2014), highlights the
increasing demand for workers with higher education levels, while the RBTC,
as explored by Fernández-Macías and Bisello (2020), classifies jobs based
on the proportion of routine or generic tasks that can be automated. Both
perspectives complement the analysis by showing how technology influences
the demand for different skills in the labor market.
4. methodoloGy Adopted to estimAte the probAbility of Job AutomAtion in
brAzil
Several studies seek to estimate the probability of automation (from now
on, p(auto)) of occupations in an economy, among which the seminal work by
Frey and Osborne (2016). They developed a methodology to estimate the
probability of automation in the set of occupations in the United States (U.S.),
based on data from the O*NET survey, the primary source of information
about occupations in that country, for the year 2010.
3 Automation could result in a “paradox of abundance”, in which society would become richer
in aggregate terms, but technological change could reinforce inequalities in power and reward for
many individuals and communities,”schema”:”https://github.com/citation-style-language/schema/raw/
master/csl-citation.json”}. (Roberts et al., 2019).
4 A panoramic reference to this ample literature that comprises both country studies and different
methodological approaches is the cross-country study from OECD (2017).
103
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
Although many studies replicate the vector of the p(auto) of occupations
estimated by Frey and Osborne (2016), labor market characteristics in the
U.S. differ from those in Latin American countries, so adjustments may
apply to conduct estimates for this region. This is what Espíndola and Suárez
(2023) did, they adapted the Frey and Osborne (2016) estimation strategy
and data source, as described ahead5, to build a vector of the probability of
automation of occupations to analyze the labor market of LA countries that
have internationally homologous occupation data (CIUO-08).
Espíndola and Suárez (2023) build on Frey and Osborne’s (2016)
categorization of occupations as automatable and non-automatable ones
but consider only two of the three original “bottlenecks to automation”: (i)
social intelligence and (ii) creative intelligence. Bottleneck (iii) perception and
manipulation was discarded based on the study by Lassébie and Quintini
(2022), who assessed the automation potential of certain activities and
occupations in light of recent advances in artificial intelligence. They found
that most skills related to the “perception and manipulation” bottleneck are
now automatable with current technologies. Therefore, Espíndola and Suárez
(2023) considered only the bottlenecks (i) and (ii), which included a set of
15 non-automatable skills that served as the main predictive variables for
automation probability, such as “collaborating with other workers”, “advising”,
and “solving complex problems”. Additionally, the authors incorporated
social indicators to enhance the predictive capacity of the algorithms by
considering important factors to better reflect task structures. Thus, it was
assumed that the variation in tasks within an occupation could be related to
the worker’s education, gender, and the sector in which they operate, among
other factors. As for the data source, Espíndola and Suárez (2023) used data
from the Programme for the International Assessment of Adult Competencies
(PIAAC) survey for four countries in Latin America (Chile, Ecuador, Mexico,
and Peru). This survey provides information about workers’ skills in their
respective countries.
It is important to mention that the vector of probability of automation of
occupations is based solely on the demand for work skills that can (or cannot)
be automated. Also, each occupation is made up of several tasks. Therefore,
the vector p(auto) for each occupation will consider how automatable the set of
activities that make up that occupation is. The vector consists of probabilities
of automation of occupations common to all workers who exercise the same
profession in any country in the region, even though the structure of tasks can
vary within the same occupation.
Even though the framework has been adapted considering the characteristics
of occupations in Latin America, it still stems from an initial structure developed
with the U.S. economy in mind. Therefore, it may not fully reflect aspects of the
Brazilian labor market, which is marked by strong informality. Furthermore, it
5 Consult Espíndola and Suárez (2023) for a complete review.
104 Kethelyn Ferreira · Marta Castilho
also does not consider people belonging to the armed forces and sex workers.
In summary, the vector should not be interpreted literally but as indicative of
underlying trends and patterns.
In this work, we will use the vector p(auto) estimated by Espíndola and
Suárez (2023) and employ the Brazilian 2022 labor market data from the
Continuous National Household Sample Survey (Pnad Contínua, according
to its Portuguese acronym) released in 2023. Notably, the results can be
different for women and men given gender differences rooted in occupational
of activities and sectoral structures. Consequently, we will analyze the data
disaggregated by sex to examine the impacts of the automation of occupations
on women. Additionally, as Pnad Contínua data allows identification of the
weight of each occupation within the economic sectors, we also present a
sectoral analysis.
5. Job AutomAtion And Gender inequAlities in brAzil
5.1. AGGreGAte And sectorAl results
In Brazil, according to the occupational structure evidenced in the job
market in 2022, there is an average probability of automation of 50%. This
probability is more pronounced for men (56%) than for women (42%).6 This
pattern is similar to the results reported by Espíndola and Suárez (2023)
for Latin American countries, where men have an average probability of
automation of 56%, while women have an average of 43%.
Conversely, Lima et al (2019) specifically estimated the probability of
automation for Brazilian jobs and found that women are in a more vulnerable
situation to automation than men, with probabilities of automation of their jobs
being 69.7% and 62.5%, respectively. It is important to note that this study
relied solely on formal employment data from the Annual Report of Social
Information(RAIS, according to its Portuguese acronym) database and used
the original automation probability vector created by Frey and Osborne
(2016). This raises concerns about the accuracy of applying the same vector
to Brazil, as the tasks performed in similar occupations can differ significantly
between the two countries.
Furthermore, we found a negative correlation between p(auto) and female
participation in occupations. In other words, the occupations with the highest
rate of female participation tend to be those with the lowest probability of
automation. However, the coefficient of determination is low; that is, the
relationship between the variables is weak (Figure 1).
In a sectoral analysis, we identified that the four sectors with the lowest
probability of automation are precisely the feminized sectors (where women
represent more than 50% of employed individuals). These sectors are domestic
services, which have a low p(auto); education; accommodation and food; and
6 Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua (2023).
105
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
health, which have a with low to medium p(auto). Conversely, the sectors with
the lowest female participation are construction, which has a medium to high
p(auto); and transport, which has a high p(auto) (Table 1).
tAble 1. probAbility of AutomAtion And pArticipAtion of men And Women by sector
Sector P(auto) Participation
Men Women
Domestic services 0.16 9% 91%
Education 0.37 26% 74%
Accommodation and Food 0.38 44% 56%
Health 0.42 26% 74%
Agriculture 0.48 80% 20%
Other services 0.49 56% 44%
Professional activities 0.53 53% 47%
R&D 0.55 59% 41%
Telecommunications and Information Services 0.55 70% 30%
Financial and Insurance Services 0.60 53% 47%
SIUP 0.64 78% 22%
Construction 0.64 96% 4%
Extractive industry 0.66 86% 14%
Transformation industry 0.69 65% 35%
Transport 0.77 88% 12%
Source: Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua
(2023). Note: P(auto) is categorized as low [0, 0.3), low-medium [0.3, 0.5), medium-high [0.5, 0.7)
and high [0.7, 1].
5.2. mAin results AccordinG to lArGe occupAtion Groups
Notably, within each of the sectors previously examined, a vast set of
activities are carried out, associated with specific occupations whose p(auto)
can also be highly dispersed. For example, in the telecommunications and
fiGure 1. distribution of employed Women’s And men’s pArticipAtion by probAbility of AutomAtion
Source: Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua
(2023).
106 Kethelyn Ferreira · Marta Castilho
information services sector, one finds directors and call-center workers among
the occupations with the greatest weight in the sector’s structure. Furthermore,
gender segregation goes beyond the concentration of women and men in
certain sectors (horizontal segregation); there is also an unequal distribution
in terms of which occupations will be performed in each sector (vertical
segregation). Therefore, understanding the potential impacts of automation on
women and men requires carrying out the analysis at the level of occupations.
First, we will conduct an analysis at the level of large groups of
occupations,7 categorized into high, medium, and low qualifications, following
the classification adopted by the International Labor Organization. In this
classification, the qualification is based on the level of skills required, which
in turn is associated with the complexity and variety of tasks demanded by
each occupation. Operationally, qualification is measured by considering the
nature of the work, the level of formal education, and the amount of informal
training or previous experience necessary for the competent performance of
tasks (ILO, 2023).
Highly qualified occupations −directors and managers, science and
intellectual professionals, and technicians and mid-level professionals− are
precisely among those with the least probability of automation. Furthermore,
these occupations are associated with better pay. However, there is a
considerable difference between the p(auto) and remuneration for technicians
and mid-level professionals concerning the other two occupations: p(auto)
tends to be higher, approaching what we consider to be a medium-high
probability of automation, and the remuneration is approximately half of that
for directors and managers (Table 2).
Together, these three occupations represent a greater weight in the female
occupational structure than in the male one; however, only science and
intellectual professionals show a higher participation of women. Furthermore,
despite the strong female participation in this occupation, it displays the
biggest salary gap for women. When analyzing director and management
positions, the concept of a “glass ceiling” is reinforced, as it prevents women
from reaching such positions (Table 2).
In the case of medium-skilled occupations, including service workers,
salespeople in shops and markets, qualified agricultural, forestry, hunting
and fishing workers, qualified workers, laborers and artisans in construction,
mechanical arts and other trades, administrative support workers, and plant
and machine operators and assemblers, the distribution of p(auto) is not
so homogeneous, varying from low-medium to high. Salaries, in turn, also
vary, despite remaining below average. Among those occupations, the only
feminized ones are service workers, salespeople in shops and markets (p(auto)
low-medium), and administrative support workers (p(auto) high). The two have,
7 According to the Classification of Occupations for Brazilian Household Surveys, which is compatible
with CIUO-08, occupations are organized into four groups: Large Group (1 digit), Main Subgroup (2
digits) Subgroup (3 digits) and Base Group (4 digits).
107
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
respectively, the first and fourth largest weight in the female occupational
structure (Table 2).
Low-skilled occupations, which include elementary ones, have a medium-
high probability of automation. In this occupational group, women represent
almost 50% of employed individuals. These occupations offer both lower
remuneration and a smaller salary gap: women, in general, earn R$1,164 per
month, which is equivalent to 89% of the remuneration received by men in the
same occupation (Table 2).
tAble 2. probAbility of AutomAtion by mAJor occupAtionAl Groups And selected chArActeristics
CIUO-08 Occupations (Large Group) P(auto) Participation Proportion Remuneration (R$)
Men Women Men Women Men Women
1 Directors and Managers 0.34 61% 39% 4% 3% 7,867 5,650
2Science and intellectuals
professionals 0.39 41% 59% 9% 17% 6,862 4,439
5Service workers, salespeople
in shops and markets 0.43 42% 58% 17% 30% 2,438 1,594
3Technicians and mid-level
professionals 0.49 56% 44% 8% 9% 3,988 2,861
9 Elementary occupations 0.53 52% 48% 15% 18% 1,303 1,164
6
Qualified agricultural,
forestry, hunting and fishing
workers
0.69 80% 20% 8% 3% 2,059 1,494
7
Qualified workers, laborers
and artisans in construction,
mechanical arts and other
trades
0.73 83% 17% 20% 5% 2,176 1,433
4Administrative support
workers 0.74 39% 61% 6% 12% 2,325 1,998
8Plant and machine operators
and assemblers 0.82 86% 14% 14% 3% 2,226 1,622
Total 0.57 57% 43% 100% 100% 2,926 2,301
Source: Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua
(2023). Note: P(auto) is categorized as low [0, 0.3), low-medium [0.3, 0.5), medium-high [0.5, 0.7)
and high [0.7, 1]. CIUO-88 groups are classified as high (1-3), medium (4-8), and low (9) qualifications.
5.3. mAin results by detAiled occupAtionAl Groups
Analyzing the main subgroups of occupations (2 digits) according to the
probability of automation, we found that women are the vast majority in
occupations with low p(auto). Of the four occupations that make up this group,
three include medium-skilled occupations and have a female participation rate
exceeding 60%. Together, they account for 20% of employed women (versus
7% in the case of men) and are commonly judged as “more feminine.” These
occupations include Health professionals; Teaching professionals; and Personal
service workers. The fourth occupation in this group is Executive directors,
public administration directors, and members of the executive and legislative
branches (Table 3).
108 Kethelyn Ferreira · Marta Castilho
Health professionals are the occupation with the highest remuneration
for men and the fifth highest for women, resulting in a salary gap exceeding
R$4,000 to the detriment of women. Among the occupations that make
up the subgroup (3 digits), women constitute the majority in nursing and
childbirth professionals, traditional and alternative medicine professionals, and
veterinarians. However, remuneration for these occupations is less than 50%
of that for doctors in general, regardless of gender. Among these occupations,
nursing and childbirth professionals have the lowest p(auto), equivalent to 0.19
(Table 3).
In the case of Teaching professionals, women only are not the majority
in the occupations of Vocational training teachers, and other music teachers
(4 digits). However, there is a notable difference in female participation in
occupations such as professors at universities and higher education teachers
(55%) compared to preschool teachers (96%). Within this group, the latter is
the occupation with the lowest p(auto), equal to 0.17, and the remuneration
for women is equivalent to only 33% of that of professors at universities and
higher education (Table 3).
Regarding personal service workers, the occupation accounts for 9% of
employed women, where they are mainly concentrated in two occupations:
cooks (3 p.p.) and Hairdressers, beauty treatment specialists and similar (5
p.p.), with these being the two with the lowest p(auto) within the group, at 0.20
and 0.18, respectively. Conversely, within the same group, the occupation
of Direct service workers for passengers can be highlighted, as its p(auto)
and remuneration are considerably above average, and the composition is
predominantly male (Table 3).
Executive directors, public administration directors, and members of the
executive and legislative branches perform highly qualified activities. These
positions have the highest paying for men and the second highest for women,
with a salary gap of R$1,658 to the detriment of women. This occupation
is commonly perceived as “more masculine,” and, indeed, women make up
only 29% of employed individuals in this field. In this group, men are the
majority regardless of the level of disaggregation (Table 3). One explanation
could be the stereotypes related to leadership qualities such as assertiveness,
rationality, risk-taking, firmness, and strategic vision, which are generally
considered masculine traits.
In the group of occupations with a low-medium probability of automation,
women are the majority in one-third of the 18 occupations that make up
the group. This group is quite diverse and includes low, medium, and high
qualification activities, which correspond to 54% and 44% of employed
women and men, respectively. Among these occupations, the two with the
greatest weight in the female occupational structure are sellers; and domestic
workers and other building interior cleaning workers (Table 3).
Even though the occupation of sellers is considered to have a low-medium
p(auto), it is among the six with the least p(auto) and includes everything
from street vendors (whether street or home-based, for example) to store
109
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
supervisors. Notably, even though salaries and p(auto) are generally low, when
analyzing the supervisory category, where salaries and p(auto) are higher, men
are the majority (Table 3).
In the case of people employed as domestic workers and other building
interior cleaning workers, the dynamics are similar. In a 4-digit analysis, there
is great granularity in the p(auto) of the occupation group. For instance, p(auto)
is 0.12 for domestic service workers in general, on the other, 0.74 for workers
cleaning the interior of buildings, offices, hotels, and other establishments, and
0.70 for vehicle washers. Female participation in these occupations is, 93%,
70%, and 8% respectively, and, for women, the average remuneration in these
occupations is R$1,034, R$1,344, and R$1,527 (compared to R$1,309,
R$1,538 and R$1,398 for men) (Table 3).
Still, regarding the group of occupations with a low to medium probability of
automation, in terms of female participation in the group of employed people,
the overrepresentation stands out in the case of personal care workers (93%)
and domestic workers and other building interior cleaning workers (80%),
considered, respectively, of medium and low qualification. Both are among the
10 lowest-paying occupations (Table 3).
On the other hand, in the highly qualified occupations that make up
the group, associated with directors and management, the maximum level
of female participation reached was 43%. Notably, female participation in
executive positions exceeding 40% only occurs when these positions are in
the hotel, food, and commercial sectors, which are the categories of directors
and management positions with the lowest remuneration. If the occupation
is associated with production and operation, the participation of women
decreases (Table 3).
In fact, within the basic occupations that make up the subgroup of directors
and production and operation managers, only managers of childcare services,
health service managers, managers of care services for elderly people, and
directors of education services have female participation exceeding 40%
(Table 3).
In the case of occupations with medium-high p(auto), women are the
majority in only one of the six occupations that make up the group, direct
customer service workers, corresponding to 73% of the employed individuals
in this occupation. This is the second lowest-paid occupation among
medium-high p(auto) occupations. This group includes telephone operators;
receptionists; and call-center workers. In a 4-digit analysis, women only do not
for the majority as hotel receptionists, where they represent 45% (Table 3).
Altogether, medium-high p(auto) occupations represent 9% of the female
occupational structure (versus 13% in the male case). This group mainly
consists of medium-skilled activities, except for garbage collectors and other
elementary occupations. In this latter occupation, women represent 22%
of employed individuals, a participation rate also found in the other basic
occupations that comprise the group (for example, garbage and recyclable
110 Kethelyn Ferreira · Marta Castilho
material collectors; and Waste classifiers), except for sweepers and similar,
where they represent 44% (Table 3).
Finally, in the group composed of occupations with high p(auto), only
medium-skilled activities are concentrated. Among them, two occupations have
greater female representation: clerks, and artisans and graphic arts workers.
Both have below-medium pay, but only the latter is among the worst-paid
occupations, behind only the skilled forestry workers, fishermen and hunters.
Considering the proportion of these occupations in the occupational structure,
the weight is relatively greater for men (35%) than for women (16%) (Table 3).
tAble 3. probAbility of AutomAtion by mAin occupAtionAl subGroup And selected chArActeristics
CIUO-08 Occupations (Main Subgroup) P(auto) Participation Proportion Remuneration
(R$)
Men Women Men Women Men Women
2 Health professionals 0.23 33% 67% 1% 3% 9,775 5,598
2 Teaching professionals 0.26 24% 76% 2% 7% 4,563 3,468
1
Executive directors, public admin-
istration directors and members
of the executive and legislative
branches
0.28 71% 29% 0% 0% 9,614 7,956
5 Personal service workers 0.28 36% 64% 4% 9% 1,948 1,536
1Managers of hotels, restaurants,
shops and other services 0.32 57% 43% 1% 1% 5,101 3,948
5 Sellers 0.32 45% 55% 10% 16% 2,612 1,715
5 Personal care workers 0.34 7% 93% 0% 5% 1,793 1,227
1Administrative and commercial
managers 0.35 60% 40% 1% 1% 9,200 6,250
1Production and operation directors
and managers 0.35 65% 35% 1% 1% 8,260 6,272
2Professionals in law, social and
cultural sciences 0.35 46% 54% 2% 3% 6,437 5,296
9Domestic workers and other build-
ing interior cleaning workers 0.36 20% 80% 3% 14% 1,467 1,131
9Itinerant service workers and
the like 0.37 63% 37% 0% 0% 1,328 999
9 Food preparation helpers 0.39 30% 70% 0% 1% 1,726 1,375
9
Elementary workers in mining,
construction, manufacturing and
transport
0.40 88% 12% 7% 1% 1,353 1,482
3Mid-level healthcare and related
professionals 0.44 28% 72% 1% 4% 3,293 2,208
6Farmers and qualified agricultural
workers 0.44 79% 21% 7% 3% 2,144 1,556
2Science and engineering profes-
sionals 0.45 67% 33% 1% 1% 7,526 4,818
3Mid-level science and engineering
professionals 0.45 83% 17% 2% 1% 3,595 2,816
3Mid-level professionals in legal,
social, cultural and related services 0.47 55% 45% 1% 1% 4,689 3,542
2Information and communications
technology professionals 0.48 77% 23% 1% 0% 7,090 5,658
111
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
CIUO-08 Occupations (Main Subgroup) P(auto) Participation Proportion Remuneration
(R$)
Men Women Men Women Men Women
6Skilled forestry workers, fishermen
and hunters 0.49 82% 18% 1% 0% 1,064 706
9Elementary agricultural, fishing and
forestry workers 0.49 80% 20% 3% 1% 1,040 953
2
Specialists in the organization
of public administration and
companies
0.50 51% 49% 2% 2% 7,019 4,563
3Mid-level professionals in financial
and administrative operations 0.54 56% 44% 3% 3% 4,654 3,514
3Mid-level information and commu-
nications technology technicians 0.59 85% 15% 1% 0% 3,301 3,394
9Garbage collectors and other
elementary occupations 0.64 78% 22% 1% 0% 1,250 1,138
4 Direct customer service workers 0.66 27% 73% 1% 4% 1,968 1,617
7
Qualified workers and workers in
metallurgy, mechanical construc-
tion and similar areas
0.68 95% 5% 6% 0% 2,478 2,521
4Numerical calculation workers and
material recorders 0.70 69% 31% 2% 1% 2,309 2,175
7Workers specializing in electricity
and electronics 0.71 95% 5% 2% 0% 2,235 2,297
7
Workers and officials in food
processing, wood, clothing and
similar areas
0.72 52% 48% 3% 4% 2,034 1,358
4 Clerks 0.76 33% 67% 2% 7% 2,466 2,176
4Other administrative support
workers 0.77 59% 41% 0% 0% 2,492 2,048
5Protection and security services
workers 0.77 89% 11% 3% 0% 2,594 2,818
7 Artisans and graphic arts workers 0.78 45% 55% 1% 1% 1,820 946
7Skilled workers and construction
workers exclusive electricians 0.80 97% 3% 9% 0% 2,037 2,304
8Operators of fixed installations and
machines 0.80 60% 40% 3% 2% 1,885 1,448
8 Assemblers 0.88 82% 18% 1% 0% 2,212 1,949
8Vehicle drivers and operators of
heavy mobile equipment 0.88 96% 4% 11% 1% 2,307 2,201
Total 0.53 57% 43% 100% 100% 2,926 2,301
Source: Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua
(2023). Note: P(auto) is categorized as low [0, 0.3), low-medium [0.3, 0.5), medium-high [0.5, 0.7)
and high [0.7, 1]. CIUO-08 groups are classified as high (1-3), medium (4-8), and low (9) qualifications.
5.4. unpAckinG the chAllenGes
In short, the weight of low and low-middle p(auto) occupations is noticeably
greater for women (74%) than for men (52%), which is reflected in a higher
aggregate p(auto) in the male case, as evidenced at the beginning of the previous
section. However, one must consider how positive this fact is.
112 Kethelyn Ferreira · Marta Castilho
On the one hand, in the case of occupations with low p(auto), the only high-
skilled occupation in the group has a female representation of less than 30% and a
negligible weight in the female occupational structure. On the other hand, although
occupations with low-medium p(auto) represent 54% of the female occupational
structure, the low-skilled occupations within this group correspond to 18pp of this
54%, while high-skilled occupations correspond to just 3pp (Table 3).
Furthermore, a disaggregated analysis allows for a more nuanced examination
of the gender pay gap, as illustrated in the box below.
box 1. Gender pAy GAp Across occupAtions
The gender pay gap is a clear and persistent reality. On the one hand, there is a significant discrepancy in the
average wages of female-dominated occupations compared to male-dominated ones. On the other hand, it is
evident that, overall, women tend to earn less than men.
At the main subgroup level (or 2 digits), we find a set of 39 occupations, of which only 12 are feminized. These
feminized occupations are, on average, precisely the ones with the lowest remuneration: the average monthly
remuneration for masculinize occupations is R$2,837, while for feminized ones is R$2,412. In other words,
remuneration in feminized occupations represents only 85% of the average remuneration in masculinized oc-
cupations.
In an analysis of the wage gap within these two groups, in feminized occupations, women tend to earn less than
70% of what men earn, while in masculinized occupations, women’s remuneration tends to be slightly higher
(equivalent to 103% of male remuneration).
In an analysis considering a more disaggregated level of occupations (4 digits), the results are similar, but
the discrepancy between men and women is accentuated. The average salary for masculinized occupations is
R$2,929, while for feminized occupations, it is R$2,251. In percentage terms, the remuneration of feminized
occupations now represents 77% of the remuneration of more masculinized occupations, compared to the 85%
it represented in a 2-digit analysis.
Regarding the salary gap, in feminized occupations, women continue to earn less than 70% of what men do.
However, unlike the previous case, the salary gap in more masculine activities continues to be to the detriment
of women, even though it is considerably smaller: women now earn 93% of men’s remuneration.
From these facts, we can conclude that men tend to work in activities considered “more prestigious” and that
pay better. In addition to a clear concentration of women in occupations that have, on average, lower pay, when
men work in these same occupations (feminized), they tend to receive considerably higher pay than women.
For example, we can consider occupations associated with domestic service, which are clearly feminized and
associated with low pay. In this case, while women receive, on average, R$1,131, men receive R$1,467. Further-
more, looking at the basic occupations that make up the group, male participation will be greater in activities
that have above-average remuneration, such as Workers cleaning the interior of buildings, offices, hotels and
other establishments, where they represent 30% of employed individuals. Conversely, their participation as
Domestic service workers in general, the occupation with the lowest pay within this group, is only 7%.
Source: Own elaboration based on data from Pnad Contínua (2023).
Another important issue is that most occupations held by women are
associated with the care economy and life reproduction tasks (with a weight
equivalent to 36% for women versus 10% for men), such as Domestic work and
health, which require cognitive skills −such as empathy, attention to the needs of
individuals who require care, communicative skills, or creativity− and thus tend
to be more difficult to automate. Additionally, these also tend to be more invisible
and poorly paid (or unpaid) activities, which discourages their automation (Table
4). As highlighted by Castilho and Ferreira (2024), 95% of employed women tend
to balance paid work with household chores and/or caregiving, whereas only 85%
of men face similar conditions. Moreover, most women dedicate more than 14
hours per week to such tasks, while most men contribute a maximum of 14 hours.
113
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
tAble 4. probAbility of AutomAtion of selected occupAtions AssociAted With the cAre economy And
selected chArActeristics
Selected Occupations (Base Group) P(auto) Participation Proportion Remuneration (R$)
Men Women Men Women Men Women
Domestic service workers in general 0.12 7% 93% 1% 9% 1,309 1,034
Child caregivers 0.18 3% 97% 0% 2% 1,598 1,009
Cooks 0.20 21% 79% 1% 3% 2,248 1,489
Manual clothes washers and ironers 0.22 30% 70% 0% 0% 1,499 1,241
Personal care workers at home 0.29 6% 94% 0% 2% 1,480 1,340
Housekeepers and domestic butlers 0.35 12% 88% 0% 0% 5,740 2,101
Community health workers 0.36 29% 71% 0% 1% 2,347 2,007
Total 0.25 11% 89% 2% 17% 2,317 1,460
Source: Own elaboration based on data from Espíndola and Suárez (2023) and Pnad Contínua (2023).
Note: P(auto) is categorized as low [0, 0.3), low-medium [0.3, 0.5), medium-high [0.5, 0.7) and high [0.7,
1].
Besides the points highlighted in sections 5.1, 5.2, and 5.3, it should be noted
that women may face greater difficulties adapting to the jobs that may arise with
the advancement technological innovations and the introduction of automation in
the labor market, in addition to taking less advantage of the potential of ICT.
According to Lawrence, Roberts, and King (2017, p. 24), “automation is more
likely to accelerate wealth and income inequalities than to create a future of mass
unemployment”, and these are shaped by people’s class, race, age, and gender
(Roberts et al., 2019). In this sense, the economic benefits of automation are likely
to be concentrated among technology and business owners, as well as highly
skilled workers, while the labor market is polarized between high-skilled and low-
skilled jobs (Lawrence et al., 2017).
Several factors contribute to this scenario, including women’s limited access
to the internet (ITU, 2021), the prevalence of a large digital gender gap (CEPAL,
2019b), the underrepresentation in STEM careers (Weller et al., 2019), and the
disproportionate burden of unpaid work and care for women (Vaca Trigo and
Valenzuela, 2022).
6. finAl considerAtions
As new technologies that encourage the automation of activities are
introduced, several authors indicate the potential for job substitution, job
precariousness, and/or job creation. These effects can occur isolated or
simultaneously, and the net effect is uncertain. Given the gender differences
rooted in both the sectoral structure of activities and the occupational
structure, these effects may vary for women and men, raising concerns about
the potential effects of the automation of occupations for women in Brazil.
Employing the vector of the probability of automation of occupations
developed by Espíndola and Suárez (2023) for Latin American countries, the
114 Kethelyn Ferreira · Marta Castilho
present study found that, in Brazil, the p(auto) in the job market for women
(42%) is lower than that for men (56%). However, one must consider the
underlying complexity: occupations with lower p(auto) tend to be lower-skilled
and associated with lower pay.
Additionally, most women’s occupations are linked to the care economy
and reproductive tasks, such as domestic work and healthcare. These activities
require specific cognitive skills, such as empathy, attention to individual needs,
and communicative or creative skills, making them less prone to automation.
However, these occupations are often informal, poorly paid, or unpaid, which
poses challenges for valuing work and disincentivizes their automation. It
is worth noting that such results should not be interpreted literally but as
indicative of underlying trends and patterns.
Furthermore, the gender pay gap is a persistent issue. In general, men
tend to work in activities considered “more prestigious” and that pay better. In
addition, when working in feminized occupations, men tend to get considerably
higher pay than women.
Finally, it should be noted that women currently face greater hardships
in entering jobs created by automation in the labor market. Factors such as
limited access to connectivity, lower digital skills, underrepresentation in STEM
careers, and the responsibility for carrying out unpaid domestic and care work
represent crucial bottlenecks for women and make it more difficult for them to
enter these potential jobs.
references
Acemoglu, D., and Autor, D. (2011). Skills, tasks and technologies: Implications
for employment and earnings. Handbook of Labor Economics, 4, 1043–
1171.
Acemoglu, D., and Restrepo, P. (2018). Artificial intelligence, automation and
work. [s.l.]: National Bureau of Economic Research, Inc.
Barrientos, S. (2001, julio 1). Gender, flexibility and global value chains.
Black, S. E., and Spitz-Oener, A. (2010). Explaining women’s success:
Technological change and the skill content of women’s work. The Review of
Economics and Statistics, 92(1), 187–194.
Boddy, D., Kearney, M., and Hershbein, B. (2015). The future of work in the age
of the machine. A Hamilton Project Framing Paper.
Brussevich, M., Dabla-Norris, E., and Khalid, S. (2019). Is technology widening
the gender gap? Automation and the future of female employment.
Brussevich, M., Dabla-Norris, E., Khalid, S., and Others. (2018). Gender,
technology, and the future of work.
Brynjolfsson, E., and McAfee, A. P. (2011). Race against the machine: How
the digital revolution is accelerating innovation, driving productivity, and
irreversibly transforming employment and the economy.
115
Job AutomAtion in brAzil: Which consequences for Women’s employment?
revistA de economíA mundiAl 69, 2025, 95-116
Brynjolfsson, E., Mitchell, T., and Rock, D. (2018). Economic consequences of
artificial intelligence and robotics: What can machines learn and what does
it mean for occupations and the economy?
Castilho, M., and Ferreira, K. (2024). Retomada industrial e emprego no Brasil:
Perspectivas de gênero e raça. In Reindustrialização Brasileira: Desafios e
oportunidades.
CEPAL. (2018). La ineficiencia de la desigualdad. Santiago: CEPAL.
CEPAL. (2019a). El comercio digital en América Latina: ¿Qué desafíos enfrentan
las empresas y cómo superarlos? [s.l.]: CEPAL.
CEPAL. (2019b). La autonomía de las mujeres en escenarios económicos
cambiantes. Santiago: CEPAL.
Davis, A. (2016). Mulheres, raça e classe. Boitempo Editorial.
Elson, D. (1999). Labor markets as gendered institutions: Equality, efficiency
and empowerment issues. World Development, 27(3), 611–627.
Espíndola, E., and Suárez, I. (2023). Automatización del trabajo y desafíos
para la inclusión laboral en América Latina: Estimaciones de riesgo
mediante aprendizaje automático ajustadas a la región. Santiago: Comisión
Económica para América Latina y el Caribe (CEPAL).
Fernández-Macías, E., and Bisello, M. (2020). A taxonomy of tasks for assessing
the impact of new technologies on work. JRC Working Papers Series on
Labour, Education and Technology. Comisión Europea.
Fontana, M. (2003). The gender effects of trade in developing countries: A
review of the literature. Discussion Papers in Economics.
Frey, C. B., and Osborne, M. A. (2016). The future of employment: How
susceptible are jobs to computerization? Technological Forecasting and
Social Change, 114, 254–280.
Goos, M., Manning, A., and Salomons, A. (2014). Explaining job polarisation:
Routine-biased technological change and offshoring. American Economic
Review, 104(8), 2509–2526.
Gonzalez, L. (1984). Racismo e sexismo na cultura brasileira. Revista Ciências
Sociais Hoje, 2(1), 223–244.
Howcroft, D., and Rubery, J. (2019). ‘Bias in, bias out’: Gender equality and the
future of work debate. Labour and Industry, 29(2), 213–227.
ILO. (2023). The International Standard Classification of Occupations (ISCO-
08) companion guide.
Lassébie, J., and Quintini, G. (2022). What skills and abilities can automation
technologies replicate and what does it mean for workers?: New evidence.
OECD Social, Employment and Migration Working Papers (Issue 282).
Lawrence, M., Roberts, C., and King, L. (2017). Managing automation:
Employment, inequality, and ethics in the digital age. IPPR.
Lerner, G. (2019). A criação do patriarcado: História da opressão das mulheres
pelos homens. São Paulo: Editora Cultrix.
Lima, Y., et al. (2019). O futuro do emprego no Brasil: Estimando o impacto da
automação. Laboratório do Futuro - UFRJ.
116 Kethelyn Ferreira · Marta Castilho
McKinsey Global Institute. (2019). The future of women at work: Transitions in
the age of automation.
Melo, H., and Castilho, M. (2009). Trabalho reprodutivo no Brasil: Quem faz?
Revista de Economia Contemporânea, 13, 135–158.
Mokyr, J., Vickers, C., and Ziebarth, N. (2015). The history of technological
anxiety and the future of economic growth: Is this time different? Journal
of Economic Perspectives.
Novick, M. (2018). El mundo del trabajo: Cambios y desafíos en materia de
inclusión. Santiago: Comisión Económica para América Latina y el Caribe
(CEPAL).
OECD. (2017). OECD Employment Outlook 2017. Paris: OECD Publishing.
Roberts, C., et al. (2019). The future is ours: Women, automation and equality
in the digital age. IPPR.
Rocha, G. R., and Vaz, D. V. (2023). Mudança tecnológica e polarização do
emprego no Brasil. Revista da ABET, 22(1).
Saffioti, H. (1987). O poder do macho. São Paulo: Moderna.
Teixeira, M. (2016). Avanços e continuidades para as mulheres no mundo do
trabalho (2004-2014). Revista da ABET.
UNESCO (2019). I’d blush if I could: Closing gender divides in digital skills
through education. UNESCO.
Vaca Trigo, I., and Valenzuela, M. E. (2022). Digitalización de las mujeres
en América Latina y el Caribe: Acción urgente para una recuperación
transformadora y con igualdad. Santiago: Comisión Económica para
América Latina y el Caribe (CEPAL).
Weller, J.; Gontero, S. y Campbell, S. Cambio tecnológico y empleo: una
perspectiva latinoamericana. Riesgos de la sustitución tecnológica del
trabajo humano y desafíos de la generación de nuevos puestos de trabajo.
Santiago: Comisión Económica para América Latina y el Caribe (CEPAL),
2019.
