Artificial Intelligence and employment: a systematic review

ACYPRESTE, RAFAEL DE; PARANÁ, EDEMILSON

doi:10.1590/0101-31572022-3320

Tabla de contenido Anterior Actual Siguiente

Tabla de contenido

Articles • Brazil. J. Polit. Econ. 42 (4) • Oct-Dec 2022 • https://doi.org/10.1590/0101-31572022-3320 linkcopy

Artificial Intelligence and employment: a systematic review

Inteligência Artificial e emprego: uma revisão sistemática

Authorship SCIMAGO INSTITUTIONS RANKINGS

This paper presents a systematic literature review, grounded on bibliometric procedures, of the (political economy) works, produced from 2008 to 2020, on the relations between Artificial Intelligence and employment. It detects a growing tendency of published papers in this field, especially from 2019, and identifies four main groups of concerns on this topic. Within these groups, a prevalence of more optimistic over skeptical accounts and, especially, of economic orthodox over heterodox approaches on the issue can be noted. Overall, it is possible to understand that both the reviewed works and their metrics are quite dispersed and varied in scope. Among other reasons, this is due to the lack of a common basic definition, within the field, of AI in the first place.

KEYWORDS:
Artificial Intelligence; automation; work.

Title	Author(s)	Contributions
Why Are There Still So Many Jobs? The History and Future of Workplace Automation	Autor, David H. (2015, 386 cit.)	Defends the idea that the past waves of innovations are not reliable parameters to analyze the new one. According to the author, there is a “polarization” of the employments, characterized by a shrinking of medium-skill jobs in comparison with the low - and high-skill jobs.
Artificial Intelligence in Service	Huang, Ming-Hui; Rust, Roland T. (2018, 135 cit.)	Develops a theory for the replacement of workers by AI in the service sector. They conclude that AI evolution accounts for a threat particularly for the jobs in the lower-intelligence categories.
Revisiting the Risk of Automation	Arntz, Melanie; Gregory, Terry; Zierahn, Ulrich (2017, 38 cit.)	Criticizes the previsions of replacement of workers for AI and robots, once they overestimate the automatable parts of jobs, disregard heterogeneity in tasks, and pay no attention to the adaptability of jobs to the digital transformations.

Title	Author(s)	Contributions
Revisiting the Risk of Automation	Arntz M., Gregory T., Zierahn U. (2017, 54 cit.)	Referred above.
Robots and Humans-Complements or Substitutes?	DeCanio S.J. (2016, 27 cit.)	The author uses Houthakker’s model to analyze the effects of AI over wages in the United States. Keeping other things equal, the wages tend to fall regardless of whether the elasticity of substitution between human labor and AI is greater than 1.9.
Automated Pastures and the Digital Divide: How Agricultural Technologies Are Shaping Labour and Rural Communities	Rotz S., Gravely E., Mosby I., Duncan E., Finnis E., Horgan M., LeBlanc J., Martin R., Neufeld H.T., Nixon A., Pant L., Shalla V., and Fraser E. (2019, 23 cit.)	Evaluate the consequences of a “digital revolution” in agriculture. They investigate the impacts on rural labor, based on interviews with locals in Canada, and conclude that there is a possibility of increasing inequalities.

Title	Author(s)	Contributions
The Future of Employment: How Susceptible Are Jobs to Computerization?	CB Frey, MA Osborne (Frey e Osborne, 2017)	They realize an analysis of susceptibility to computerization of 702 occupations in the United States. For the authors, 47% of the jobs are in the high-probability category of automation, with 70% or more of automation risk.
Trends in US Wage Inequality: Revising the Revisionists	DH Autor, LF Katz, MS Kearney (2008, 2,696 cit.)	The paper examines the growth of wage inequality in the United States since 1963, which presents a pattern variation in 1980. Looking for a complete analysis, they study the role of information technologies in wage changes. In conclusion, the authors theorize about the “polarization” of the job market.
The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market	H David, D Dorn (2012, 2,500 cit.)	Having the paper by Autor, Katz, and Kearney (2008, above) as a starting point, the authors add a model of spatial equilibrium to evaluate the impacts over salaries. They obtained evidence that the effects of “polarization” guide workers from medium skills and routine tasks to the service sector, less subject to robotization.
Job Polarization in Europe	M Goos, A Manning, A Salomons (2009, 1,380 cit.)	The authors estimate the job “polarization” between 1993 and 2006 in 16 European countries caused by the routine-task automation allowed by the recent technological advances.
The Risk of Automation for Jobs in OECD Countries	M Arntz, T Gregory, U Zierahn (2016, 1,200 cit.)	Evaluates the chances of future reduction of jobs as a result of automation and digitalization. However, the authors argue this as unexpected regardless of whether the analysis is conducted based on tasks and not on the occupations, which have a high level of aggregation.
Robots and Jobs: Evidence from US Labor Markets	D Acemoglu, P Restrepo (2017, 967 cit.)	Analyzes the growth of the use of “industrial robots” in previous human-executed tasks in the United States between 1990 and 2007. According to the authors’ estimation, one robot more per thousand workers may reduce the employment-population ratio by 0.18%-0.34% and wages by 0.25%-0.50%.

Title	Author(s)	Contributions
Industria 4.0: fabricando el futuro	A Inés-Basco, G Beliz, D Coatz, P Garnero (2018, 39 cit.)	Analyzes the impact of the new technologies on manufacturing, workplace, and global chains of value. In addition, it presents some evaluations about Argentina and Brazil.
Las transformaciones tecnológicas y su impacto en los mercados laborales	J Weller (2017, 20 cit.)	Evaluates the current transformations in the labor market and argues that the impacts of the technical change depend largely on policy choices made by the relevant players.
Economía de plataformas y empleo:¿ Cómo es trabajar para una app en Argentina?	J Madariaga, C Buenadicha, E Molina, C Ernst (Madariaga et al., 2019)	The aim is to classify and analyze the platform work in Argentina. Some alleviations of unemployment were identified as well as challenges for labor regulation and worker protection.
Nuevas tendencias en el empleo: retos y opciones para las regulaciones y políticas del mercado de trabajo	G Bensusán (2016, 17 cit.)	A discussion about the expansion of precarious jobs, especially in Latin America. Furthermore, the report outlines some public policy to deal with this question.
Desempleo tecnologico: una aproximacion al caso latinoamericano	A Aguilera, MG Ramos Barrera (Aguilera e Ramos Barrera, 2016)	The main goal is to analyze the relations between science and technology, innovation, and unemployment rate in seven Latin America countries (1996-2011). The paper concludes that there is no evidence, until now, that technology investments reached the level to increase unemployment.
Las transformaciones tecnológicas y sus desafíos para el empleo, las relaciones laborales y la identificación de la demanda de cualificaciones	G Bensusán Areous, W Eichhorst, JM Rodríguez (2017)	The book gathers three studies related to technological transformations and labor market. They are broadly interested in the education processes and productivity questions as well as their implications in the economics of developing countries.
Na era das máquinas, o emprego é de quem? Estimação da probabilidade de automação de ocupações no Brasil	PH Albuquerque, CAPB Saavedra, RL de Morais (2019, 6 cit.)	The report estimates the automation probabilities of occupations in Brazil, applying the same methodology used by Frey & Osborne (2017). High and very high automation probabilities are found in 54.45% of the formal workers.

location_on

Centro de Economia Política Nove de Julho Avenue, 2029, CEP 01313-902 - São Paulo - SP - Brazil, Tel. (55 11) 3816-6053 - São Paulo - SP - Brazil
E-mail: bjpe@fgv.br

rss_feed Acompañe los números de esta revista en su lector de RSS

[TFN1] Source: WoS. Elaborated by the authors.