Algorithmic human resource management: Synthesizing developments and cross-disciplinary insights on digital HRM

Classified as probably not a substantial claim.

Algorithmic HRM is a specific subset/type of digital HRM.

“Using this definition, we regard algorithmic HRM as a specific subset/type of digital, or electronic, HRM (Bondarouk et al., 2017;Strohmeier, 2020b).”

• cites Bondarouk et al., 2017 sourcePartly supported

  The Introduction treats electronic HRM as a broad category encompassing digital forms and HRM content shared through IT. It does not identify algorithmic HRM as a subset; that specific classification is the citing authors’ inference, not a point stated in this source. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Strohmeier, 2020b sourceNot checked — no full text of the source

May combine two claims.

Software algorithms support HRM decision making in two ways: through providing information (or augmentation) and with automation.

“Software algorithms support HRM decision making in two ways, through providing information (or augmentation) and with automation (Leicht-Deobald et al., 2019;Raisch & Krakowski, in press).”

• cites Leicht-Deobald et al., 2019 sourceSupported
• cites Raisch & Krakowski, in pressNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Predictive algorithms operate regression-based forecasting techniques that help managers predict which employees are likely to leave the organization, or to predict the future performance of a job candidate.

“These so-called predictive algorithms operate regression-based forecasting techniques that, for example, help managers predict which employees are likely to leave the organization (and thus how to make decisions about retention), or to predict the future performance of a job candidate (and thus help hiring managers with selection decisions) (Cheng & Hackett, 2021;Leicht-Deobald et al., 2019).”

• cites Cheng & Hackett, 2021 sourceNot checked — no full text of the source
• cites Leicht-Deobald et al., 2019 sourcePartly supported

  In “A Brief History of Algorithm-Based HR Decision-Making” (page unavailable), the paper says predictive algorithms use advanced regression techniques and describes a hiring algorithm that scores how well applicants would fit a job. It does not describe using predictive algorithms to identify employees likely to leave or to guide retention decisions; its mention of turnover concerns a descriptive scorecard. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim.

Algorithms, human agency and social processes in OLPs are likely to differ due to their business model and how they interact with (platform) workers.

“algorithms, human agency and social processes in OLPs are likely to differ due to their business model and how they interact with (platform) workers (Gegenhuber et al., 2020;Kuhn & Maleki, 2017;Meijerink & Keegan, 2019;Wood et al., 2019).”

• cites Gegenhuber et al., 2020 sourceNot checked — no full text of the source
• cites Kuhn & Maleki, 2017 sourceNot checked — no full text of the source
• cites Meijerink & Keegan, 2019 sourceCould not verify

  The available indexed passages are bibliographic metadata (some labeled Abstract; no page numbers), not discussion of platform practices. Searches for algorithms, human agency, business models, and interactions with platform workers found no passage addressing the claim. (LLM, conf=0.72)

  Experimental — accuracy not yet measured

May combine two claims.

OLPs substitute algorithmic HRM for organizational hierarchy and physical oversight to control worker behavior, and need to strike a balance between autonomy and control.

“For instance, rather than use organizational hierarchy to control employee behavior through physical oversight, OLPs substitute algorithmic HRM and in doing so, need to strike a balance between autonomy and control (Frenken et al., 2020;Meijerink et al., 2021;Wood et al., 2019).”

• cites Frenken et al., 2020 sourceNot checked — no full text of the source
• cites Meijerink et al., 2021 sourcePartly supported

  The Introduction describes algorithm-based compensation and other HRM practices that control gig workers, while ‘Institutional complexity of HRM for freelancers’ describes the tension between worker autonomy and control. The paper does not say that algorithmic HRM substitutes for organizational hierarchy or physical oversight; its discussion of the corporation logic instead likens platform control to organizational routines and hierarchy. Page numbers are unavailable in our copy. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Wood et al., 2019 sourcePartly supported

  The Findings and discussion section says direct supervision of remote gig workers was limited by distance, while platform ratings and rankings provided effective algorithmic control and left workers substantial discretion. It also suggests that extinguishing that discretion would undermine service quality. The paper does not explicitly say platforms need to strike a balance between autonomy and control, so that prescriptive part goes beyond its findings. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

May combine two claims.

OLPs rely on opaque HRM algorithms in an attempt to replace HR managers and disguise the control exercised over the behavior of their freelance workforce.

“Moreover, OLPs rely on opaque HRM algorithms in an attempt to replace HR managers and disguise the control that is exercised over the behavior of their freelance workforce (Meijerink et al., 2021).”

• cites Meijerink et al., 2021 sourcePartly supported

  The Abstract and Implications for theory and practice describe technology-enabled, covert HRM practices that control freelance workers while preserving their freelance status. Limitations and directions for future research says Uber Eats and Deliveroo withheld details of their algorithms. The copy does not say that the algorithms are intended to replace HR managers or that their opacity itself disguises control. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

In contrast with 'traditional' contexts, platform workers are not supervised by human (line) managers tasked with implementing HRM activities or HR specialists on HR analytics teams.

“In contrast with 'traditional' contexts, platform workers are not supervised by human (line) managers tasked with implementing HRM activities (Bos-Nehles et al., 2013) or HR specialists on HR analytics teams (Ellmer & Reichl, this issue; Wiblen & Marler, this issue).”

• cites Bos-Nehles et al., 2013 sourceSupported
• cites Ellmer & Reichl, this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Wiblen & Marler, this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

The analysis highlights the epistemic practices of boundary spanning, customizing dashboards, and speaking a language of numbers.

“Their analysis highlights the epistemic practices of boundary spanning, customizing dashboards, and speaking a language of numbers.”

• cites Ellmer & Reichel, this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Gig workers who work with online labor platforms spend a lot of time trying to converse with humans as well as figuring out how to challenge, resist or game the controls and constraints built into the technology.

“Gig workers who work with online labor platforms spend a lot of time trying to converse with humans as well as figuring out how to challenge, resist or game the controls and constraints built into the technology.”

• cites Waldkirch et al., this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

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

HR managers see and use HRM algorithms to demonstrate their quantitative and business knowledge, and therefore enhance their credibility and status within the organization.

“They show how HR managers see and use HRM algorithms to demonstrate their quantitative and business knowledge, and therefore enhance their credibility and status within the organization.”

• cites Ellmer & Reichel, this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

From these data, the authors find that HRM activities such as selection, performance management, compensation and employee relations are automated by HRM algorithms.

“From these data, the authors find that HRM activities such as selection, performance management, compensation and employee relations are automated by HRM algorithms.”

• cites Waldkirch et al., this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Taken together, these practices reflect what the authors refer to as epistemic alignment, where members in an HR analytics team align algorithmic output with the way organizational decision-makers perceive business reality.

“Taken together, these practices reflect what the authors refer to as epistemic alignment, where members in an HR analytics team align algorithmic output with the way organizational decision-makers perceive business reality.”

• cites Ellmer & Reichel, this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

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

The final paper provides an example of online labor platforms as a context where algorithmic HRM fully automates HRM decision-making.

“The final paper in this special issue provides an example of online labor platforms as a context where algorithmic HRM fully automates HRM decision-making.”

• cites Waldkirch et al., this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Digitization by means of sensors and smart devices allows for the real-time collection of digital data that come in large amounts and from a variety of sources.

“Digitization by means of sensors and smart devices allows for the real-time collection of digital data that come in large amounts and from a variety of sources.”

May combine two claims.

Studies into HR analytics, HRM algorithms, and AI deployed in HRM practices tend to use different terms to describe seemingly similar content, leading to a lack of construct clarity that may prevent the scholarly community from building a collective and coherent body of knowledge.

“Studies into HR analytics ( Marler & Boudreau, 2017;Minbaeva, 2017;Tursunbayeva et al., 2018;Van den Heuvel & Bondarouk et al., 2017 ), HRM algorithms (Cheng & Hackett, 2021;Leicht-Deobald et al., 2019), and artificial intelligence (AI) deployed in HRM practices (Strohmeier & Piazza, 2015;Vrontis et al., 2021), while beginning to coalesce around key issues, tend to use different terms to describe”

• cites Suddaby, 2010 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Algorithmic HRM requires digitization-a process by which information is turned into digital data.

“As such, algorithmic HRM requires what (Strohmeier, 2020b) refers to as digitization -a process by which information is turned into digital data.”

• cites Strohmeier, 2020b sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Automation of HRM-related decision-making distinguishes algorithmic HRM from other forms of digital HRM.

“Accordingly, its first two features (i.e.digital data and use of software algorithms) imply that algorithmic HRM falls within the realm of digital HRM, while the third feature, automation of HRM-related decision-making, distinguishes algorithmic HRM from other forms of digital HRM.”

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

Algorithms process data and by so doing they produce 'information' that can be treated as new knowledge or not.

“Accordingly, algorithms process data and by so doing they produce”

• cites Pachidi et al., 2021Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Software algorithms are (or can be made) opaque to workers, and it may not be entirely clear how these algorithms learn, or the exact source of their learning, and how they add parameters or adjust parameter weights using the data to which they are dynamically adapting.

“For example, software algorithms are (or can be made) opaque to workers, and it may not be entirely clear how these algorithms learn, or the exact source of their learning, and how they add parameters or adjust parameter weights using the data to which they are dynamically adapting (Burrell, 2016).”

• cites Burrell, 2016 sourceNot checked — no full text of the source

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

The use of AI in HRM is a specific form of algorithmic HRM.

“In this sense, the use of AI in HRM is a specific form of algorithmic HRM.”

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

Software algorithms transform data in an automated and efficient manner.

“This is where software algorithms come into play for transforming data in an automated and efficient manner.”

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

For software algorithms to use data, these data need to be presented as binary digits that a computer can read and process.

“For software algorithms to be able to”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Given the sheer size and speed at which they are collected, Big Data are difficult to process efficiently in a manual manner.

“Given the sheer size and speed at which they are collected, Big Data are difficult to process efficiently in a manual manner.”

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

Research on the ethical issues of HR analytics is still at an early stage and focused mainly on ethical risks such as discrimination, bias, privacy concerns and excessive surveillance.

“As noted by Tursunbayeva et al. (2021), research on the ethical issues of HR analytics -where HRM algorithms are mostly used for augmenting decision making -is still at an early stage and focused mainly on ethical risks such as discrimination, bias, privacy concerns and excessive surveillance (Tursunbayeva et al., 2021).”

• cites Tursunbayeva et al., 2021Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Organizations that 'become digital' do not need to pay attention to balancing autonomy-control tensions that OLPs incorporate into their HRM algorithms.

“As such, these organizations do not need to pay attention to balancing autonomy-control tensions that OLPs incorporate into their HRM algorithms.”

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

While almost all organizations today are part of the information technology revolution, most organizations are still in a digitalization process, albeit some at a more advanced stage than others.

“While almost all organizations today are part of the information technology revolution, most organizations are still in a digitalization process, albeit some at a more advanced stage than others (Strohmeier, 2020b).”

• cites Strohmeier, 2020b sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Employees may not be directly affected by algorithmic HRM in cases where a human manager and HRM algorithms both interact to enact HRM activities, yet where humans make the final decision.

“Moreover, employees may not be directly affected by algorithmic HRM in cases where a human manager and HRM algorithms both interact to enact HRM activities, yet where humans make the final decision.”

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

Actions of workers generate digital data that HRM algorithms process, thereby triggering changes in the materiality of algorithmic HRM.

“Vice versa, the actions of workers generate digital data that HRM algorithms processes, thereby triggering changes in the materiality of algorithmic HRM.”

Classified as probably not a substantial claim.

Algorithmic HRM includes the use of a broad array of digital HRM activities that now goes well beyond the use of descriptive/predictive algorithms that HR analytics scholars frequently study.

“As these examples show, algorithmic HRM includes the use of a broad array of digital HRM activities that now goes well beyond the use of descriptive/predictive algorithms that HR analytics scholars frequently study.”

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

Seen through a socio-material lens, HRM algorithms are performative while triggering changes in social processes.

“Seen through a socio-material lens, HRM algorithms are performative while triggering changes in social processes.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Given the size of most workforces operating on OLPs, the execution of HRM activities by human managers would be too time-consuming and limit platforms' capacity to scale rapidly and perform well.

“Given the size of most workforces operating on OLPs, the execution of HRM activities by human managers would simply be too time-consuming and limit platforms' capacity to scale rapidly and perform well.”

May combine two claims.

A recent survey of almost 9,000 business and HR leaders shows that the majority of organizations mostly collect worker data for descriptive purposes regarding turnover (82%), salary costs (68%) or workforce composition (53%), while predictive and prescriptive HRM algorithms are rarely used.

“For instance, a recent survey of almost 9,000 business and HR leaders shows that the majority of the organizations mostly collect worker data for descriptive purposes regarding turnover (82%), salary costs (68%) or workforce composition (53%), while predictive and prescriptive HRM algorithms are rarely used (Deloitte, 2020).”

• cites Deloitte, 2020Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The gig workers on the Upwork platform are independent contractors and along with its completely digital business model, most of the HRM practices are automated.

“The gig workers on the Upwork platform are independent contractors and along with its completely digital business model, most of the HRM practices are automated.”

• cites Waldkirch et al., this issueNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Prescriptive algorithms that automate HRM activities not only forecast what can be done in different scenarios, but also select and execute a course of action without (much) involvement by a human decision maker.

“Prescriptive algorithms that automate HRM activities not only forecast what can be done in different scenarios, but also select and execute a course of action without (much) involvement by a human decision maker.”

May combine two claims.

OLPs work with freelancers whose alternative employment relationship is temporary and transactional, in contrast to most existing organizations where work is performed by employees in open-ended employment relationships.

“In contrast to most existing organizations in which work is performed by employees who exchange their labor as part of an open-ended employment relationship, OLPs work with freelancers whose alternative employment relationship is temporary and transactional (Prassl, 2018).”

• cites Prassl, 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Although the workforce that performs tasks through OLPs generally consists of freelance workers, OLPs also have employees (e.g., software engineers and data specialists) who are responsible for designing the algorithmic HRM activities that automate HRM decisions regarding platform workers.

“Although the workforce that performs tasks through OLPs generally consists of freelance workers, OLPs also have employees, for example software engineers and data specialists, who are responsible for designing the algorithmic HRM activities that automate HRM decisions regarding platform workers (Kuhn, Meijerink & Keegan, 2021).”

• cites Kuhn, Meijerink & Keegan, 2021 sourceNot checked — no full text of the source

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

HR algorithms based on artificial intelligence and machine learning render the performance evaluation of freelancers completely opaque and prevent them from autonomously managing or marketing their capabilities.

“In this special issue, Waldkirch et al. describe how HR algorithms based on artificial intelligence and machine learning render the performance evaluation of freelancers completely opaque and prevent them from autonomously managing or marketing their capabilities.”

• cites Waldkirch et al.Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S4 (based).

With increasingly sophisticated algorithms, particularly with advances in AI algorithms, some HR decisions are now made without human 'decision-making'.

“But with increasingly sophisticated algorithms, particularly with advances in AI algorithms, some HR decisions are now made without human”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Labor sociologists have shown that the use of prescriptive HRM algorithms enables employers to capture excessive economic value from workers' labor.

“Labor sociologists have shown how the use of prescriptive HRM algorithms enables employers to capture excessive economic value from workers' labor.”

May combine two claims.

HRM algorithms can both automate and augment HRM decision making.

“Put differently, HRM algorithms can both automate and augment HRM decision making.”

Probably not a substantial claim: may not pass S4 (based).

Organizations that 'become digital' are more likely to have conventional employment contracts along with managerial oversight as the primary control mechanism.

“In contrast, organizations that 'become digital' are more likely to have conventional employment contracts along with managerial oversights that serves that the primary control mechanism.”

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

Human agency and social processes in OLPs are likely to differ due to their business model and how they interact with (platform) workers.

“Besides differing in the material agency of HRM algorithms, human agency and social processes in OLPs are likely to differ due to their business model and how they interact with (platform) workers”