Cited by 118Digital Economy and Work TransformationEthics and Social Impacts of AILegal and Labor Studies

Artificial intelligence and work: a critical review of recent research from the social sciences

Open access
Deranty, Jean-Philippe, Corbin, Thomas·2024·AI & SOCIETY

Abstract

This review seeks to present a comprehensive picture of recent discussions in the social sciences of the anticipated impact of AI on the world of work. Issues covered include: technological unemployment, algorithmic management, platform work and the politics of AI work. The review identifies the major disciplinary and methodological perspectives on AI’s impact on work, and the obstacles they face in making predictions. Two parameters influencing the development and deployment of AI in the economy are highlighted: the capitalist imperative and nationalistic pressures.

AI snapshot

Written from the paper’s full text.

TL;DR

This critical review synthesises recent social-science research on AI's impact on work across four areas: technological unemployment, algorithmic management, platform work, and the politics of AI work. It argues that AI's effects are not technologically determined but shaped by two background imperatives—a capitalist profit imperative and nationalistic/geopolitical pressures.

Key findings

  • Predictions of technological unemployment vary enormously by method: Frey and Osborne estimate 47% of US jobs at risk within two decades, Bowles 54% in the EU and USA, while Arntz et al. (using an alternative task-linking method) find around 9% across the OECD (12% in Austria, 6% in South Korea), and Nedelkoska and Quintini estimate 14% of jobs highly automatable and 32% at 50–70% risk.
  • Optimistic accounts report productivity and job gains: Accenture estimates AI could raise labour productivity by up to 40% by 2035; PwC estimates global GDP up 14% by 2030 (up to $15.7 trillion); the World Economic Forum predicts a net increase of 58 million jobs (133 million new roles, 75 million displaced); McKinsey values AI in banking at $1 trillion annually (15.4% of sales).
  • The review identifies methodological weaknesses in forecasting: models rely on existing job/task descriptions, treat "skill" and "routine work" as unproblematic, and cannot anticipate new tasks and jobs; Clifton et al. are cited that technology's employment impact "is not deterministic."
  • Marxist and historical critiques argue capitalist property relations and profit criteria, weakened labour power, cheap labour, shareholder-value priorities, and sectoral overcapacity all disincentivise job-replacing automation, so absence of mass unemployment coincides instead with precariousness and exploitation.

Methods

  • Design: a critical, non-disciplinary literature review, one output of an Australian Research Council–funded project (DP190103116) on the centrality of work; "critical" is defined via the "Macbeth question" about the social-political context of knowledge claims.
  • Scope: work is restricted to formal economic activity (production of goods and services for profit or wage); AI is defined as computational methods using data to replicate activities in labour processes.
  • Analysis: thematic synthesis across four areas—technological unemployment (Sect. 3), algorithmic management (Sect. 4), platform work (Sect. 5), and the politics of AI work (Sect. 6)—organised around the capitalist and nationalistic imperatives.

Limitations

  • Stated by the authors: the lack of a specific disciplinary perspective and the broad thematic scope carry drawbacks, including the risk of missing important details and important references in each disciplinary field; they also note their restricted definition of work excludes social reproduction activities.
  • Evident from the text: no systematic search or inclusion protocol is described, and the review relies on secondary literature and cited estimates rather than primary data collection; the authors themselves note that case-study methods inform only a particular job at a particular time and place, and that generalising across national contexts is difficult.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s00146-022-01496-x

Cites

161 of 161 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 137 cited sources checked · 0 supported · 0 flagged

  • Interpretationnone given

    The development and deployment of AI in the economy are influenced by two parameters: a capitalist imperative and nationalistic pressures.

    “Two parameters influencing the development and deployment of AI in the economy are highlighted: the capitalist imperative and nationalistic pressures.”
    • Also stated in: § 1 Introduction · plain
      “AI processes are deployed in a context characterised by two fairly uncontroversial features: a capitalistic imperative and nationalistic pressures.”
    Quote in the paperSubstantial claim
  • Interpretationcitation

    The geopolitical factor is directly tied to the economic one.

    “The geopolitical factor is directly tied to the economic one.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hwang 2018

      Not checked in the free run.

  • Interpretationreasoning

    Advances in theoretical knowledge and technical improvements alone are not sufficient to explain the actual paths digital innovation travels and the forms it takes.

    “yet it is not sufficient to refer to advances in theoretical knowledge and technical improvements alone to explain the actual paths it travels and the forms it takes.”
    Quote in the paperSubstantial claim
  • Premisecitation

    In recent years, the battle for geopolitical hegemony has meant that other major powers, notably China, have also invested significantly in AI research.

    “In more recent years, the battle for geopolitical hegemony has meant that other major powers, notably China, have also invested significantly in AI research (Barton et al. 2017;Allen 2019;Savage 2020;Roberts et al. 2021).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Barton et al. 2017

      Not checked in the free run.

    • cites Allen 2019

      Not checked in the free run.

    • cites Savage 2020 source

      Not checked in the free run.

  • Premisecitation

    A major driver of AI innovation and development since the 1980s in the US has been the military.

    “A major driver of AI innovation and development since the 1980s in the US has been the military (Berman 1992;Morgan et al. 2020 on US investment in AI for intel-ligence and surveilance systems).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Berman 1992 source

      Not checked in the free run.

    • cites Morgan et al. 2020

      Not checked in the free run.

  • Classified as probably not a substantial claim. 61
  • May combine two claims. 50
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Crowdwork and app-work are so far not widespread employment models.

    “or crowdwork and app-work become widespread employment models (so far they are not)”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim.

    Premisecitation

    AI works by processing large amounts of data to identify patterns.

    “AI works by processing large amounts of data to identify patterns (Van Rijmenam 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Van Rijmenam 2019 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    AI development is stoked by an alliance of corporate and military interests: economic competition is one aspect of the battle for geopolitical hegemony, and military supremacy serves to ensure economic prosperity in competition with other nations.

    “AI development is stoked by an alliance of corporate and military interests: economic competition is one aspect of the battle for geopolitical hegemony, and military supremacy serves to ensure economic prosperity in competition with other nations (Hwang 2018 for the geopolitics of controlling semiconductors manufacturing and AI patents).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hwang 2018

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Previous waves of automation propelled by advances in information and communication technology were biased in favor of workers with higher skills, replacing lower skill workers and assisting workers with pre-existing complex skill sets.

    “It is well established that previous waves of automation propelled by advances in information and communication technology were biased in favor of workers with higher skills, replacing lower skill workers and assisting workers with pre-existing complex skill sets (Acemoglu and Autor 2011, Göranzon and Josefson 2012, Buera et al. 2015, Mellacher and Scheuer 2020).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Acemoglu and Autor 2011 source

      Not checked in the free run.

    • cites Göranzon and Josefson 2012

      Not checked in the free run.

    • cites Buera et al. 2015

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    AI is introduced in workplaces for increased efficiency in some technical aspect of the work process, or explicitly to replace human workers and reduce labour costs.

    “AI is introduced in workplaces for increased efficiency in some technical aspect of the work process, or it is introduced explicitly with the aim of replacing human workers and thereby reduce labour costs.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Premisecitation

    AI has benefited from investment by companies hoping to cut labour and transaction costs, increase outputs, raise productivity, manage workforce, refine customer knowledge, and establish monopoly positions.

    “AI has benefited from investment by companies that hope to find in it a new avenue for cutting costs, notably labour (Bessen et al. 2018) and transaction costs (Gurkaynak 2019; Lobel 2018), increasing outputs through rationalization of the production process (Acemoglu and Restrepo 2019), raising productivity (Brynjolfsson et al. 2019), managing the workforce in more efficient ways (Eubanks 2022),”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Bessen et al. 2018

      Not checked in the free run.

    • cites Gurkaynak 2019

      Not checked in the free run.

    • cites Acemoglu and Restrepo 2019 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisenone given

    Digital innovation, like all modern technology, has its own momentum.

    “Digital innovation, like all modern technology, certainly has its own momentum, yet it is not sufficient to refer to advances in theoretical knowledge and technical improvements alone to explain the actual paths it travels and the forms it takes.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    AI did not emerge and is not being deployed in an economic or political void.

    “This means that AI did not emerge and is not being deployed in an economic or a political void.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    A 2020 McKinsey report evaluates the annual value increase of AI in the banking industry at $1 trillion, or 15.4% of sales.

    “A 2020 McKinsey reports evaluates the annual value increase of AI in the banking industry at $1 trillion, or 15.4% of sales.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The history of automation to date demonstrates that new technologies so far have consistently created new jobs, both directly and indirectly.

    “The history of automation to date demonstrates that new technologies so far have consistently created new jobs, both directly and indirectly (David 1990).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites David 1990

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The creation of new tasks is a key component of the labour economists' arguments against universal technological unemployment.

    “The creation of new tasks is a key component of the labour economists' arguments against universal technological unemployment.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    Comparable innovation effects are expected in other industries where the innovation process is inseparable from the research process.

    “Comparable innovation effects are expected in other industries where the innovation process is inseparable from the research process (Cockburn et al. 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Cockburn et al. 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Frey and Osborne predict that 47% of all jobs within the U.S. are at risk of technological replacement within two decades.

    “The two business scholars famously predict that 47% of all jobs within the U.S. are at risk of technological replacement within two decades.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Frey and Osborne 2017 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Ernst et al. (2019) conclude that AI is likely to play out in ways different from previous waves, by increasing productivity and potentially creating more inclusive growth, provided correct educational measures are taken in countries that are part of globalised labour chains.

    “In an important review of the literature, which combines a historical approach with a task-focused one, Ernst et al. (2019) conclude that AI in fact is likely to play out in ways different from previous waves, by increasing productivity and potentially creating more inclusive growth, provided correct educational measures are taken in countries that are part of globalised labour chains.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Ernst et al. 2019 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    A 2018 World Economic Forum study predicts that automation will result in a net increase of 58 million jobs, with 133 million new roles created and 75 million current workers displaced.

    “A 2018 study by the World Economic Forum (2018) predicts that automation will result in a net increase of 58 million jobs, with a total 133 million new roles created and 75 million current workers displaced.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites World Economic Forum 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    AI is viewed by many business experts as a technology that can spur innovation and provide competitive advantage.

    “AI is viewed by many business experts as a technology that can spur innovation and provide competitive advantage (Davenport 2018; Polak et al. 2020 for the finance sector).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Davenport 2018

      Not checked in the free run.

    • cites Polak et al. 2020 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Frey and Osborne's predictions were refuted by Arntz et al. (2016), who found much lower outcomes, around 9% across the OECD, with a high of 12% in Austria and as low as 6% in South Korea.

    “Using this approach, Frey and Osborne's earth-shaking predictions on the impact of AI was refuted by Arntz et al. (2016).”
    • Also stated in: § 3.2.2 Optimistic scenarios · plain
      “With this new method, they found evidence for much lower outcomes, around 9% across the OECD (a high of 12% in Austria) and as low as 6% in South Korea.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Arntz et al. 2016

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Another trend widely anticipated is the polarisation of labour markets, similar to what occurred in the previous wave of automation.

    “Beside displacement, another trend widely anticipated is the polarisation of labour markets (Brito 2020;Korinek and Stiglitz 2019), similar to what occurred in the previous wave of automation”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Brito 2020

      Not checked in the free run.

    • cites Korinek and Stiglitz 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Bowles claims that 54% of jobs in the EU and USA were under threat in the same time span.

    “Using a similar approach, Bowles arrived at an even higher figure, claiming that 54% of jobs in the EU and USA were under threat in the same time span (Bowles 2014).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bowles 2014

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    A key Manpower survey in 2017 showed that managers in some countries expected to substitute workers for machines, while others expected AI to increase hiring; overall the survey demonstrated optimism.

    “A key Manpower survey in 2017 (ManpowerGroup 2017) provided interesting figures, with managers in some countries expecting to substitute workers for machines, whilst others expected AI to increase hiring.Overall, the survey demonstrated optimism (also ServiceNow 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites ManpowerGroup 2017

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Gillham et al. estimate that global GDP would increase by 14% by 2030, an equivalent of up to $15.7 trillion.

    “Gillham et al. (2018) in a PWC report estimate that global GDP would increase by 14% by 2030, an equivalent of up to $15.7 trillion.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Gillham et al. 2018

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Felten et al. developed a model for how computerisation and specific AI functions can affect activities making up occupations, and they land on more measured conclusions.

    “In another noteworthy approach, Princeton computer expert Edward Felten and his colleagues developed a model for how computerisation and specific AI functions can affect the activities making up particular occupations.They also land on more measured conclusions (Felten et al. 2018(Felten et al. , 2019(Felten et al. , 2021)).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Felten et al. 2018

      Not checked in the free run.

    • cites Felten et al. 2019

      Not checked in the free run.

    • cites Felten et al. 2021

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Human skills can be complemented rather than copied by automated processes, with machines taking charge of routine aspects of the job.

    “Finally, human skills can be complemented rather than copied by automated processes, with machines taking charge of the routine aspects of the job (Autor 2014;Brooks et al. 2020;Alarie et al. 2018;Ekbia and Nardi 2017).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Autor 2014

      Not checked in the free run.

    • cites Brooks et al. 2020 source

      Not checked in the free run.

    • cites Alarie et al. 2018 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

    Premisecitation

    Many economists and technology experts contend that AI will substitute for human work at such a scale that social-economic organisations will be shaken to the ground.

    “Many economists and technology experts contend that AI will substitute for human work at such a scale that socialeconomic organisations will be shaken to the ground as a result.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Danaher 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Purdy and Daugherty estimate that AI has the potential to increase labor productivity across the board by up to 40% in 2035.

    “Purdy and Daugherty in a 2016 Accenture report (Accenture 2016) estimate that AI has the potential to increase labor productivity across the board by up to 40% in 2035.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Accenture 2016

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Many economists, historians, business scholars and executives reject pessimistic visions of massive technological unemployment.

    “Many economists, historians, business scholars and executives reject such pessimistic visions of massive technological unemployment (Moghaddam et al. 2020).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Moghaddam et al. 2020

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Another macro-economic argument highlighted by Autor (2015) is captured in the image of the O-ring.

    “Another macro-economic argument highlighted by Autor (2015) is captured in the image of the O-ring.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Autor 2015

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Bughin (2020) concludes that labor redistribution will occur.

    “Via a questionnaire targeting 3000 companies, Bughin (2020) concludes that labor redistribution will occur.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bughin 2020

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The ILO has published many reports on the future of work, which tend to be cautiously optimistic.

    “The ILO has published many reports on the future of work, which tend to be cautiously optimistic (for instance ILO 2018).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites ILO 2018

      Not checked in the free run.

  • May combine two claims.

    Premisenone given

    In the pharmaceutical industry, AI is already used to spur innovation in mapping rare diseases, predicting pharmaceutical properties, and identifying new therapeutic uses.

    “In the pharmaceutical industry, AI is already used to spur innovation in a number of ways: for instance, to cite just three platforms, by mapping rare diseases (BERG platform); predicting the pharmaceutical properties of small-molecule candidates for drug development (XtalPi's ID4 platform); or identifying new therapeutic uses for already validated products (BioXel).”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Nedelkoska and Quintini estimate that 14% of jobs are highly automatable (probability over 70%) and 32% have a risk between 50 and 70%.

    “In a multi-country approach covering 32 nations, Nedelkoska and Quintini (2018) estimate that 14% of jobs are highly automatable (probability of automation over 70%), 32% have a risk of between 50 and 70%.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Nedelkoska and Quintini 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Acemoglu and Restrepo summarise four logics associated with the productivity effect: (1) rise in demand for labor follows from economic growth triggered by innovation; (2) demand for labour can increase because automation triggers increased demand for capital; (3) automation can deepen existing automation, increasing productivity without substituting labour; (4) AI automation creates new tasks.

    “In their contribution to The Economics of Artificial Intelligence (2017) Acemoglu and Restrepo summarise in plain terms the complex logics associated with the productivity effect.First, as noted, there is a rise in the demand for labor that follows automatically from economic growth triggered by innovation.In theory the demand for labour can be witnessed even in sectors where automation occurred.”
    • Also stated in: § 3.2.2 Optimistic scenarios · plain
      “Second, demand for labour can increase because automation triggers increased demand for capital.Third, automation can deepen existing automation, which increases productivity without substituting labour (since it is machines that are improved).And fourthly, and most importantly, AI automation creates new tasks.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Acemoglu and Restrepo 2017

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Productivity has been stagnant in developed economies over the past seventy years, with growth per decade decreasing from 2.3% in the 1950s to 1.8% in the 2010s.

    “It is a well-known fact that productivity has been stagnant in developed economies over the past seventy years, with growth per decade decreasing from 2.3% in the 1950s, to 1.8% in the 2010s (Gries and Naudé 2018).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Gries and Naudé 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    New technology creates entirely new jobs.

    “New technology creates entirely new jobs (Wilson et al. 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Wilson et al. 2017

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    A major use of technology is to bypass labour forces when they are well organized and push back effectively against capital imperatives.

    “A major use of technology is to bypass labour forces when they are well organized and push back effectively against capital imperatives (Mueller 2021).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Mueller 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    At the macro-level, the factors involved in economic reality across different contexts are so numerous and variable that significant unpredictability in how AI might complement rather than replace human work should be assumed.

    “At the macro-level, the factors involved in economic reality across different context are so numerous and variable, one would assume significant unpredictability in how AI might complement rather than replace human work.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    On some accounts, this multiplier effect might even be felt in manufacturing industries servicing automated factories.

    “On some accounts, this effect might even be felt in manufacturing industries servicing automated factories (Goos et al. 2015).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Goos et al. 2015

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Strict property rules mean that technological innovation is exclusively owned, put to use, and the outcomes of its productivity-raising potentials captured, by private owners.

    “Strict property rules mean that technological innovation is exclusively owned, put to use, and the outcomes of its productivity-raising potentials captured, by private owners.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Nieswandt 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    In a world prone to climate crisis, the environmental impact of AI might well constrain its deployment, at least if sustainability becomes a serious parameter in economic activity.

    “Similarly, in a world prone to climate crisis, the environmental impact of AI (Corbett 2018;Dhar 2020;Lucivero 2020;Strubell et al. 2020; van Wynsberghe 2021) might well constrain its deployment, at least if sustainability becomes a serious parameter in economic activity (Nishant et al. 2020 for a contrary optimistic view).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Corbett 2018 source

      Not checked in the free run.

    • cites Dhar 2020 source

      Not checked in the free run.

    • cites Lucivero 2020 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationnone given

    Methods that rely on case studies only inform on a particular job or occupation at a particular time and place, and there are methodological risks in generalising from particular cases.

    “Similarly, methods that rely on case studies only inform on a particular job or occupation at a particular time and place, and there are methodological risks in generalising from particular cases.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim.

    Premisecitation

    The capitalist imperative means that the deployment of technologies in a capitalist context is guided exclusively by the search for profit.

    “The capitalist imperative means that the deployment of technologies in a capitalist context is guided exclusively by the search for profit.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Nieswandt 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    The absence of technological unemployment is nothing to celebrate because it coincides with higher levels of precariousness, underemployment and exploitation.

    “The absence of technological unemployment is in itself nothing to celebrate though, as it coincides with higher levels of precariousness, underemployment and exploitation.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Even in the last decade, productivity growth slowed significantly.

    “Even in the last decade, productivity growth slowed significantly (Brynjolfsson et al. 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Brynjolfsson et al. 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Surplus value can be extracted just as well from human workers.

    “Surplus value can be extracted just as well from human workers (Dinerstein 2021).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dinerstein 2021 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Private property in the means of production and the profit motive make it unlikely that AI-based automation will lead to massive job losses.

    “Drawing on Cohen's classical reconstruction of historical materialism (2000), Barbara Nieswandt (2021) shows that private property in the means of production and the profit motive make it unlikely that AI-based automation will lead to massive job losses.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Nieswandt 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    Many jobs are actually more difficult to perform than is often assumed.

    “This is true not only for technical reasons, because many jobs are actually more difficult to perform than is often assumed, but also for economic ones: as De Stefano makes the case (2020), assuming that automation necessarily increases productivity might in some cases be based on an overly narrow view of the tasks automated and on erroneous measures of efficiency.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    The models appear to describe mathematically consistent worlds, but it is less clear that they describe our messy world or the future.

    “The models appear to describe mathematically consistent worlds, but it is less clear that they describe our messy one, let alone what it might be in the future.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Overcapacity in manufacturing and agriculture leads to an exodus of workers towards service and care sectors, where productivity is achieved through wage suppression and depreciation of working conditions.

    “Overcapacity in manufacturing and agriculture leads to an exodus of workers towards service and care sectors, where productivity is achieved through wage suppression and depreciation of working conditions.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Benanav 2020

      Not checked in the free run.

    • cites Smith 2020

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    The impact of technology on employment is not deterministic; the deployment of new technologies is contingent upon a multitude of factors, including public policy, firm strategy and geography.

    “an important review by Clifton et al. (2020, 11) highlights that, “the impact of technology on employment is not deterministic-the deployment of these new technologies is contingent upon a multitude of factors, including public policy, firm strategy and geography, among others.””
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Clifton et al. 2020

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Methods used to predict labour-market trajectories deliver substantial lessons for understanding past trends but are far less convincing for future ones.

    “For the external observer, these methods seem to deliver substantial lessons for understanding past trends, but are far less convincing when it comes to future ones.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The frequency with which mathematical models are revised, with new parameters and axiomatic hypotheses in each new paper, creates a gap between the assurance of conclusions stated in plain English and the ability of the models to capture reality.

    “The frequency with which the mathematical models are revised, with new parameters and axiomatic hypotheses being introduced in each new paper, gives the non-spe-cialist the sense that there is a gap between the assurance with which conclusions are stated in plain English and the ability of the models to capture reality, notably given the number of idealizing assumptions underpinning the analyses.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Labour's current deficiency in power compared to capital's means that automation is bounded.

    “Labour's current deficiency in power compared to capital's means that automation is “bounded”.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Fleming 2019 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    AI needs data regarding workers' skill, time use, and behavior, which in turn makes worker monitoring a necessity.

    “AI needs data regarding workers' skill, time use, and behavior, which in turn makes worker monitoring a necessity.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Interpretationreasoning

    These new occupations come with low wages, precarious working conditions and tenure, and uncertain hours.

    “These new occupations come with low wages, precarious working conditions and tenure and uncertain hours.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Given the other tools capitalists possess to ensure the exploitation of workers, the search for profit is better served by increasing output than job churn, so there is no incentive to shed jobs.

    “Given the other tools capitalists possess to ensure the exploitation of workers, the search for profit is better served by increasing output than job churn, which means that there is no incentive to shed jobs.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    When human labour is relatively cheap, there is no incentive for capital to invest and deploy expensive technologies.

    “When human labour is relatively cheap, there is no incentive for capital to invest and deploy expensive technologies.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    Automation triggers employment in new service sectors, even a return to older forms of dependent labour.

    “This triggers employment in new service sectors, even a return to older forms of dependent labour.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    In the O-ring model, failure of any one step in the chain of production leads the entire production process to fail; improvements in reliability of any link increase the value of improvements in all others.

    “In the O-ring model, failure of any one step in the chain of production leads the entire production process to fail.Conversely, improvements in the reliability of any given link increase the value of improvements in all of the others.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Kremer 1993

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    If needs, tasks and jobs will be new, it is difficult to guess what they might be, let alone capture them in models premised on existing job profiles and descriptions of tasks in currently existing occupations.

    “By definition, the lay reader is tempted to say, if needs, tasks and jobs will be new, it seems difficult to guess what they might be, let alone capture them in models premised on existing job profiles, and on the descriptions of the tasks entailed in currently existing occupations (see the 2018 ILO literature review for precisely this point; also Arntz et al. 2019).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites ILO 2018

      Not checked in the free run.

    • cites Arntz et al. 2019

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Marx-inspired analyses of technological unemployment are supported by historians who contradict optimistic readings like Mokyr's and emphasise the slowing down of innovation under financial capitalism, where profit maximisation occurs through speculation rather than changes in industrial paradigms.

    “These Marx-inspired analyses of technological unemployment are supported by the analyses of historians who directly contradict optimistic readings like Mokyr's, and emphasise the slowing down of innovation under financial capitalism, where profit maximisation occurs through speculation rather than changes in industrial paradigms (Gordon 2014(Gordon , 2015)).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Gordon 2014

      Not checked in the free run.

    • cites Gordon 2015 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Algorithmic management raises issues of privacy, not just at the workplace but also at home, notably following the pandemic-induced shift to home-based working.

    “Surveillance and control: algorithmic management raises obvious issues of privacy (Bhave et al. 2020;Ebert et al. 2021;Fukumura et al. 2021;Tsamados et al. 2022), not just at the workplace, but also at home, notably following the pandemic-induced shift to home-based working (Collins 2020).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Bhave et al. 2020

      Not checked in the free run.

    • cites Ebert et al. 2021 source

      Not checked in the free run.

    • cites Collins 2020

      Not checked in the free run.

  • May combine two claims.

    Interpretationnone given

    AI gathers and processes data in ways that are often hidden from workers, leading to decisions made in ways which humans cannot, or at least cannot quickly or efficiently, process.

    “AI also gathers and processes data in ways that are often hidden from workers, leading to decisions made in ways which humans cannot, or at least cannot quickly or efficiently, process.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The power of AI for gathering and processing vast amounts of data can be harnessed in traditional work settings where employees are hired under work arrangements and labour contracts predating AI automation.

    “The power of AI for gathering and processing vast amounts of data can be harnessed in traditional work settings where employees are hired under work arrangements and labour contracts predating AI automation.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    What can in some conditions amount to an increase in autonomy can also mean a lack of training, notably around OHS issues.

    “But what can in some conditions amount to an increase in autonomy can also mean a lack of training, notably around OHS issues.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    The computational power of AI allows it to function not just as a new industrial tool within pre-existing work processes, but also as the centerpiece of a new business model that radically alters modes of working, conditions of employment, and interactions of workers with management and customers.

    “However, the computational power of AI allows it to function not just as a new industrial tool within pre-existing work processes.It also becomes the centerpiece of a new business model that radically alters modes of working, as well as the conditions of employment and the interactions of workers with management and customers.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Premisecitation

    Other battles are about the right to privacy of workers and the development of regulatory frameworks to give an ethical frame to AI development.

    “Other battles are about the right to privacy of workers (De Stefano 2020 for Europe) and the development of regulatory frameworks to give an ethical frame to AI development (Salento 2018).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites De Stefano 2020

      Not checked in the free run.

    • cites Salento 2018 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Vallas and Schor conclude that power asymmetry rather than horizontal democracy has so far been the experience of workers.

    “Vallas and Schor reference a large number of studies on working conditions for appworkers (see the more recent Moore and Woodstock 2021), and conclude that power asymmetry rather than horizontal democracy has so far been the experience of workers.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Vallas and Schor 2020 source

      Not checked in the free run.

    • cites Moore and Woodstock 2021 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Workers are managed and directed through reasons they are not provided with, preventing collaborative or even consultative leadership styles and substituting a directive or even coercive one in its place.

    “Workers are managed and directed through reasons they are not provided with, preventing collaborative or even consultative leadership styles, substituting a directive or even coercive one in its place (Dunphy and Stace 1993).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Dunphy and Stace 1993 source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Many platform workers use the work to complement income, and precarisation through platform work is far from a universal trend.

    “Vallas and Schor reject an assumed view of platform workers that is overly homogeneous.Many platform workers in fact use the work to complement income, precarisation through platform is far from a universal trend.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Vallas and Schor 2020 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Surveillance can lead to increased pressure on workers to perform, taking away moments of respite, as is well documented in warehouse and platform work.

    “Surveillance can lead to increased pressure on workers to perform, taking away moments of respite, as is well documented in warehouse (Hanley and Hubbard 2020) and platform work (Newlands 2021a, b).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Hanley and Hubbard 2020

      Not checked in the free run.

    • cites Newlands 2021a, b

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    A lot of crowdwork and gigwork is highly flexible, has non-regular hours, which can have a negative impact on work-life balance and affect physical and mental health and personal relations.

    “A lot of crowdwork and gigwork is highly flexible, has non-regular hours, which can have a negative impact on work-life balance and affect physical and mental health and personal relations as a result (Muntaner 2018;Allan et al. 2021).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Muntaner 2018 source

      Not checked in the free run.

    • cites Allan et al. 2021 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The immense gap between ideal-scenario policy propositions and economic and political reality points to the need for another kind of approach, one that focuses on what makes AI developments doubtful, at least in the short term.

    “Ideal-scenario policy recommendations might have their usefulness, but the immense gap between the propositions and economic and political reality points to the need for another kind of approach, one that focuses precisely on what makes these developments of AI doubtful, at least in the short term.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationnone given

    Privacy infringements can occur at all stages of the data cycle: at the time of collection, in the analysis of the data, in the use of the data, and when data ought to be erased.

    “Privacy infringements can occur at all stages of the data cycle: at the time of collection, in the analysis of the data, in the use of the data, and when data ought to be erased.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Management by AI is by definition about removing the human element from the decision-making process.

    “Management by AI is by definition about removing the human element from the decision-making process.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    Service work for an Uber driver is far less scripted than service work in more traditional settings.

    “Service work for an Uber driver for instance is far less scripted than service work in more traditional settings.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    The company appears to be absolved of its responsibilities, removing existing avenues through which workers understand and demand these responsibilities are met.

    “The company appears to be absolved of its responsibilities, removing existing avenues through which workers understand and demand these responsibilities are met.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    Susskind's proposal is detailed and seems consistent, but it is utterly unrealistic in current ideological and geopolitical conditions, particularly in the few countries spearheading AI innovation.

    “As a blueprint, Susskind's proposal is detailed and seems consistent, but it is utterly unrealistic in current ideological and geopolitical conditions, particularly in the few countries spearheading AI innovation.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Flexible hours might also be viewed as increased autonomy.

    “Flexible hours might also be viewed as increased autonomy.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationcitation

    If AI deployment causes significant technological unemployment, wealth polarisation, or widespread crowdwork and app-work, bringing further precariousness in employment conditions, then the impact on current social organisations will be significant.

    “If, as a result of AI deployment in work, technological unemployment does occur at a significant scale, or wealth polarisation, or crowdwork and app-work become widespread employment models (so far they are not), bringing further precariousness in employment conditions, then the impact on current social organisations will be significant (Acemoglu and Restrepo 2020b).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Acemoglu and Restrepo 2020b

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Breaches of privacy touch on a fundamental human right and also represent a strong leverage tool for managers, with which they can exert control and undermine autonomy.

    “Breaches of privacy touch on a fundamental human right, but they also represent a strong leverage tool for managers, with which they can exert control and undermine autonomy (Shapiro 2018).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Shapiro 2018 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Different types of platform workers operate under different conditions and have vastly different experiences.

    “Different types of workers operate under different conditions and have vastly different experiences.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Current social systems continue to be organised on the basis of full-time employment in jobs attracting good wages as the condition for the complete enjoyment of social and economic rights, including full health and pension coverage.

    “This is because current social systems continue to be organised on the basis of full-time employment in jobs attracting good wages as the condition for the complete enjoyment of social and economic rights, including full health and pension coverage.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The view of platforms as a new form of Weberian iron cage underestimates the capacity of platform workers to evade control, resist and organise.

    “However, this view of platforms underestimates the capacity of platform workers to evade control, resist and organise.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationnone given

    What is at stake is the clash between the two imperatives noted at the outset and the plausibility of a 'good AI', or 'AI for good'.

    “What is at stake is the clash between the two imperatives noted at the outset and the plausibility of a “good AI”, or “AI for good”.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    The heterogeneity of the workforce exists also in the locations of workers, who are scattered around regions and even around the world, creating isolation and preventing organisation for collective action.

    “The heterogeneity of the workforce exists also in the locations of the workers, who are scattered around regions, and even around the world.This creates isolation and again prevents organisation for collective action.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationnone given

    Micro-tasking crowdwork goes to the heart of AI.

    “One aspect of crowdwork that is worth highlighting is micro-tasking crowdwork that goes to the heart of AI.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationcitation

    It is impossible to make overly general statements about platform work.

    “A key lesson from the sociology of work perspective is that it is impossible to make overly general statements about platform work (Schor et al. 2020).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Schor et al. 2020 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    One feature of the relaxation of management control is the heterogeneity of the workforce, which acts against organisation.

    “One feature of this relaxation of management control, is the heterogeneity of the workforce this produces, which acts against organisation.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Premisenone given

    One major battle concerns the legal status of gig workers, whether platforms can divest themselves of all responsibilities regarding the people they employ.

    “One major battle concerns the legal status of gig workers, whether platforms can divest themselves of all responsibilities regarding the people they employ.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    AI can also complement work in some professions, notably by taking away the routine aspects of the job, allowing workers to focus on the more creative or rewarding parts of the work.

    “By the same token, however, AI can also complement work in some professions, notably by taking away the routine aspects of the job.Workers then focus on the more creative, or rewarding parts of the work.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Given the rhetoric used by the main actors, it is difficult to see ethical considerations having much sway for some of the possible extensions of AI.

    “Given the rhetoric used by the main actors, it is difficult to see ethical considerations having much sway for some of the possible extensions of AI.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    One might be skeptical of the willingness of states to enforce principles of 'good AI' when they are engaged in high stakes contests over geopolitical hegemony.

    “Regarding the nationalist imperative, one might be skeptical of the willingness of states to enforce principles of “good AI” when they are engaged in high stakes contests over geopolitical hegemony.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    AI does not by itself determine 'good' nor 'bad' outcomes for the world of work; rather, what matters is the kind of world in which AI will be developed and deployed.

    “What these extraneous, economic and political, parameters indicate is that AI does not by itself determine 'good' nor 'bad' outcomes for the world of work.Rather, what matters is the kind of world in which AI will be developed and deployed.”
    Quote in the paperNot a substantial claim
93 of 93
  • The authors state this as an assumption.

    Scientific inquiries are inextricably tied to the historical context in which they occur.

    “The core assumption this question captures is that scientific inquiries are inextricably tied to the historical context in which they occur.”
  • The paper relies on this without support.

    Knowledge claims reflect social and political fault lines, whether researchers are aware of it or not.

    “Whether they are aware of it or not, knowledge claims reflect social and political fault lines.”
2 of 2

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Researchers in the field of AI and work, in their optimistic scenarios, expect a shift in tasks, the creation of new tasks and an emergence of new work profiles rather than the replacement of workers.

    “They share this expectation with researchers in the field of AI and work, who in their optimistic scenarios expect a shift in tasks, the creation of new tasks and an emergence of new work profiles rather than the replacement of workers (Deranty and Corbin 2022).”

    From Friend or foe? Exploring the implications of large language models on the science system (2025) · cites this paper as Deranty and Corbin 2022

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Section roadmap is shifted by one section

    Minor

    The roadmap places technological unemployment in Section 2, but the paper labels it Section 3. It likewise places algorithmic management and platform work in Sections 3 and 4, while their headings label them Sections 4 and 5.

  • Technological-unemployment subsection references use the wrong numbers

    Minor

    The introduction to Section 3 points readers to subsections 2.1 and 2.2, but the relevant subsections are numbered 3.1 and 3.2.

2 of 2

Cites

Flagged first, then by citations

11 of the 172 references couldn’t be matched to a known paper.

Research works printed but not linked: 40

  1. Acemoglu D (Ed.) (2021) Redesigning AI. MIT Press
  2. Amorim H, Moda F (2020) Work by app: Algorithmic management and working conditions of Uber drivers in Brazil. Work Organ Labour Global 14(1):101–118
  3. Applebaum HA (1992) The Concept of Work: Ancient, medieval, and modern. SUNY Press
  4. Arntz M, Gregory T, Zierahn U (2019) Digitization and the future of work: Macroeconomic consequences. Handbook Labor Hum Resour Popul Econ 1:1–29
  5. Autor DH, Handel MJ (2013) Putting tasks to the test: Human capital, job tasks, and wages. J Law Econ 31(S1):S59–S96
  6. Bales RA, Stone K (2020) The invisible web at work: artificial intelligence and electronic surveillance in the workplace. Berkeley J Employm Labor Law 41:1
  7. Bastani A (2018) Fully Automated Luxury Communism. Verso
  8. Berger T, Chen C, Frey C B (2017) Cities, Industrialization and Job Creation: Evidence from Emerging Countries, pp 1–25. Mimeo, Oxford Martin School
  9. Bhave DP, Teo LH, Dalal RS (2020) Privacy at work: A review and a research agenda for a contested terrain. J Manag 46(1):127–164
  10. Brito DL (2020) Automation Does Not Kill Jobs. Rice University, It Increases Inequality
  11. Brynolfsson E, McAfee A (2011) Race against the Machine. Mass: Digital Frontier Press, Lexington
  12. Clifton J, Glasmeier A, Gray M (2020) When machines think for us: the consequences for work and place. Oxford University Press UK 13:3–23
  13. Datta A, Tschantz MC, Datta A (2015) Automated experiments on ad privacy settings. Proc Priv Enh Technol 2015(1):92–112
  14. Daugherty PR, Wilson HJ (2018) Human+ machine: Reimagining work in the age of AI. Harvard Business Press
  15. David PA (1990) The dynamo and the computer: an historical perspective on the modern productivity paradox. Am Econ Rev 80(2):355–361
  16. De Stefano V (2019) Negotiating the Algorithm: Automation, Artificial Intelligence, and Labor Protection. Comp Lab L & Pol'y J 41:15
  17. Eubanks B (2022) Artificial Intelligence for HR: Use AI to Support and Develop a Successful Workforce. Kogan Page, United Kingdom
  18. Ford M (2015) Rise of the Robots. Cornell University Press, New York
  19. Ford (2021) Rule of the robots: how artificial intelligence will transform everything. UK: John Murray Press
  20. Göranzon B, Josefson I (2012). Knowledge, skill and artificial intelligence, Springer Science & Business Media
  21. Halal W, Kolber J, Davies O, Global T (2017) Forecasts of AI and future jobs in 2030: Muddling through likely, with two alternative scenarios. J Futur Stud 21(2):83–96
  22. Kalleberg AL, Vallas SP (2018) Probing precarious work: Theory, research, and politics. Res Sociol Work 31(1):1–30
  23. Lewis P, Bell K (2019) Understanding the UK’s productivity problems: new technological solutions or a case for the renewal of old institutions?. Empl Relations 41(2):296–312. https://doi.org/10.1108/ER10-2018-0273 · DOI
  24. Luksha P et al (2015) Atlas of Emerging Jobs. Skolkovo, Moscow
  25. Manyika J, Chui M, Miremadi M, Bughin J, George K, Willmott P, Dewhurst M (2017) A future that works: AI, automation, employment, and productivity. McKinsey Global Institute Research. Tech Rep 60:1–135
  26. Moghaddam Y, Yurko H, Demirkan H, Tymann N, Rayes A (2020) The future of work: how artificial intelligence can augment human capabilities, business expert press
  27. Möhlmann M, Henfridsson O (2019) What people hate about being managed by algorithms, according to a study of Uber drivers. Harvard Business Review, 30
  28. Muller Z (2019) Algorithmic harms to workers in the platform economy: The case of Uber. Columbia J Law Soc Prob 53:167
  29. Munoz JM, Naqvi A (2018) Business strategy in the artificial intelligence economy. Business Expert Press
  30. Neufeind M, O'Reilly J, Ranft F (2018) Work in the Digital Age: Challenges of the Fourth Industrial Revolution, Rowman & Littlefield International
  31. Paschkewitz J, Patt D (2020) Can AI make your job more interesting? Issues Sci Technol 37(1):74–78
  32. Purcell C, Brook P (2020) At least I’m my own boss! Explaining consent, coercion and resistance in platform work. Employment and Society, Work
  33. Scarpetta S (2018) The future of work: Advancing labor market resilience. J Int Aff 72(1):51–57
  34. Schmidt FA (2017) Digital labour markets in the platform economy. Mapp Polit Challeng Crowd Work Gig Work 7:2016
  35. Solow R (1987) "We'd better watch out", New York Times Book Review, July 12, 1987
  36. Srnicek N (2017) Platform Capitam. Polity Press, Germany
  37. Srnicek N, Williams A (2016) Inventing the Future (revised and Updated Edition): Postcapitalism and a World Without Work. Verso Books, Spain
  38. Sundararajan A (2017) The sharing economy: The end of employment and the rise of crowd-based capitalism. MIT press
  39. Susskind D (2020) A world without work: Technology, automation and how we should respond, Penguin UK
  40. Wilson HJ, Daugherty P, Bianzino N (2017) The jobs that artificial intelligence will create. MIT Sloan Manag Rev 58(4):14

Cited by

Sorted by citations

Authors
Deranty, J.-P., & Corbin, T.
Affiliations
Macquarie University (AU)
Corresponding author
  • Deranty, Jean-Philippe
  • Corbin, Thomas
Venue
AI & SOCIETY
Published
2022-06-28
Volume
39
Issue
2
Pages
675-691
ISSN
  • 0951-5666 (ISSN-L)
  • 1435-5655
OpenAlex
W4283645384
Cited by
118
References
172
Citations by year
2026: 33 · 2025: 39 · 2024: 31 · 2023: 14 · 2022: 1
FWCI
30.63
Type
journal article
Language
en
Open access
Locations
AI & Society (publishedVersion, cc-by)
Funders
  • Australian Research Council (DP190103116)
  • Macquarie University
Topics
  • Social Sciences › Sociology and Political Science › Digital Economy and Work Transformation
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Social Sciences › Business, Management and Accounting › Industrial relations › Legal and Labor Studies
Keywords
artificial intelligence, future of work, technological unemployment, algorithmic management, platform work, politics of work, capitalism, nationalism

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-07.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-07.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review