Article
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Online labor platforms (OLPs) can use algorithms along two dimensions: matching and control. While previous research has paid considerable attention to how OLPs optimize matching and accommodate market needs, OLPs can also employ algorithms to monitor and tightly control platform work. In this paper, we examine the nature of platform work on OLPs, and the role of algorithmic management in organizing how such work is conducted. Using a qualitative study of Uber drivers’ perceptions, supplemented by interviews with Uber executives and engineers, we present a grounded theory that captures algorithmic management of work on OLPs. In the context of both algorithmic matching and algorithmic control, platform workers experience tensions relating to execution, compensation, and belonging. We show that these tensions trigger market-like and organization-like response behaviors by platform workers. Our research contributes to the emerging literature on OLPs.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Algorithmic matching in platform works is the "algorithmically mediated coordination of interactions between demand and supply" (Möhlmann et al., 2021). Platforms use algorithms to perform the function of a marketplace (Griesbach et al., 2019), making information available on the platforms to facilitate decision-making and recommending the best possible matches that balance the interests of both the provider and the user-customer. ...
... Besides preliminary inputs (client's requests, time availability, locations, etc.), output data like ratings and online customer evaluation have been reported to create a salient impact on the likelihood of job matches. In the case of ride-hailing, favorable rating scores are manifested in relatively more ride allocations (Möhlmann et al., 2021), while high ranking and positive feedback lead to a higher possibility of being visible on search results to clients in crowd-work platforms (Stark & Pais, 2020). ...
... Algorithmic controlling characteristic is a crucial topic that draws the most attention in scholarly research, which can be noticed by the number of objects sorted into that category. According to Möhlmann et al. (2021), algorithmic control is the use of algorithms to supervise workers' performance and ensures it aligns with the organization's goals. Likewise, Tomprou & Lee (2022) define it as the automation of managerial practices, which were traditionally the responsibility of middle or upper management. ...
Conference Paper
Full-text available
We review the literature on algorithmic management to help future researchers acquire a comprehensive "recap" of past research with detailed discussions on the main findings and develop a taxonomy as a tool of summarization that assists researchers in reflecting critically on their systems and identifying potential gaps. We determine five critical areas of algorithmic management: the mechanisms of algorithmic management, effects of algorithmic management, second party's response to algorithmic management, concerns around algorithmic management, design of algorithmic management, and policy implications. These topics are analyzed and discussed.
... The case study by Moehlmann et al. (2020) with the Uber platform demonstrates how rules and standardized processes are built into algorithms by the platform owner to actively influence drivers' behavior. One Uber employee described how rules and algorithms contingently influence drivers: ...
... The case study with the Uber platform carried out by Moehlmann et al. (2020) provides evidence for matching facilitated through algorithmic data analysis. For example, one Uber executive explained the significance of algorithms when matching riders and drivers on their platform: ...
... The case study by Moehlmann et al. (2020) illustrates social feedback on the Uber platform. They found that although drivers on the Uber platform enjoyed their autonomy of being their own boss, they also liked being part of a community of drivers. ...
Thesis
Digital platforms consist of technical elements such as software and hardware and associated social elements such as organizational processes and standards. When such social or technical elements seem logical individually but inconsistent when juxtaposed they form tensions. Prior research on platforms often focused on individual elements of digital platforms but neglected possible related and conflicting elements which offers limited insight about underlying tensions. While some studies on platforms considered tensions, they largely assumed that centralized platform owners being responsible for responding to tensions, neglecting collective response mechanisms in blockchain-based decentralized autonomous organizations (DAOs) where decentralized participants typically respond to tensions. The examination of tensions in the context of centralized platforms and decentralized autonomous organizations offers an opportunity to surface conflicting elements that form novel types of socio-technical tensions which require collective and technology-enabled response mechanisms. This thesis explored what tensions exist in centralized and decentralized digital platform contexts and how platform participants can respond to selected tensions. For this purpose, this thesis comprises five essays that employ multiple different research methods including interviews analyzed by using techniques of grounded theory, qualitative meta-analysis of published case studies, and systematic literature reviews. The findings derived from all five essays contribute to a better understanding of tensions in digital platforms. In particular, this thesis (1) offers a lens for analyzing platforms as collective organizations in which tensions arise at the collective meta-organizational level requiring collective responses, (2) identifies new tensions and response mechanisms related to generativity and collectivity, and (3) points to a novel category of socio-technical tensions that are especially salient in digital platforms.
... In the information systems (IS) literature, algorithmic management has been defined as ''the large-scale collection and use of data on a platform to develop and improve learning algorithms that carry out coordination and control functions traditionally performed by managers'' (Möhlmann et al. , p. 2001. As such, a key distinguishing feature of algorithmic management (vis-à-vis traditional management approaches) is the use of increasingly intelligent algorithms in conjunction with digital technologies (e.g., mobile apps and sensors embedded in smartphones) not only to support, or ''informate'' (Zuboff 1985), but to automate the execution of coordination and control tasks with little to no human involvement (Cram and Wiener 2020;Möhlmann et al. 2021). The intelligence of these algorithms is largely driven by advanced technological affordances such as context-awareness, real-time responsiveness, interactivity, and (big) data availability (Kellogg et al. 2020;Schuetz and Venkatesh 2020). ...
... In any case, the actual enactment of relevant management mechanisms and their delivery/communication to workers is fully automated by algorithms and digital technology (see Fig. 1). Möhlmann et al. (2021) conceptualize algorithmic management in terms of two main dimensions: algorithmic matching and algorithmic control. Uber, for example, employs AI-based algorithms for both matching drivers with customers (including dynamic pricing) and controlling (i.e., directing, evaluating, and rewarding/sanctioning) drivers' work behaviors. ...
... Finally, algorithmic management may also interfere with the social fabric of organizations, potentially hampering the satisfaction of relatedness needs. Already, the dispersed workforce participating on ride-sharing platforms became a prime example of workers' alienation in a new platform economy (e.g., Möhlmann et al. 2021). In a similar vein, visibility of one's performance and a related scoring system may trigger competitive behaviors among workers (Levy 2015), which over time may undermine collaboration and workplace climate. ...
... Yet, current studies have mainly emphasized the negative impact of using platforms, showing it as impersonal and giving minimal support (Waldkirch et al., 2021). By heavily relying on automated processes and algorithms measuring and monitoring work activities, platforms often neglect freelancer participation or voice (Duggan et al., 2020;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2020;Waldkirch et al., 2021). Additionally, studies frequently illustrate the lack of support workers experience, for example when they encounter difficulties with clients (Brawley & Pury, 2016;Fieseler et al., 2019). ...
... Freelancers argue that they have no chance of getting in-depth insight into their work activities, platform use and the reasons for success or failure (Bush & Balven, 2021;Gegenhuber et al., 2021;Nemkova et al., 2019;Schroeder et al., 2021). While most platforms track behaviour in detail (Möhlmann et al., 2020;Waldkirch et al., 2021), such information is not provided to freelancers; rather, it is used internally to optimize algorithms for matching and control (Möhlmann et al., 2020;Norlander, Jukic, Varma, & Nestorov, 2021). Our findings offer a new facet to this discussion and advocate that displaying in-depth insights on work activities is a key for activating self-leadership among freelancers. ...
... Freelancers argue that they have no chance of getting in-depth insight into their work activities, platform use and the reasons for success or failure (Bush & Balven, 2021;Gegenhuber et al., 2021;Nemkova et al., 2019;Schroeder et al., 2021). While most platforms track behaviour in detail (Möhlmann et al., 2020;Waldkirch et al., 2021), such information is not provided to freelancers; rather, it is used internally to optimize algorithms for matching and control (Möhlmann et al., 2020;Norlander, Jukic, Varma, & Nestorov, 2021). Our findings offer a new facet to this discussion and advocate that displaying in-depth insights on work activities is a key for activating self-leadership among freelancers. ...
... Digital sensors and algorithmic decision-making are altering the way that organizations set work rules, monitor worker activities, provide feedback to workers, and enforce compliance [23,64]. This approach, termed algorithmic control, enables organizations to guide and influence day-to-day worker behavior using technology rather than human managers [71]. However, by doing so, organizations are faced with uncertainty, not only in terms of how effectively worker behavior is influenced by the digital algorithms, but also in terms of the corresponding impact on worker well-being (e.g., stress, overload, enjoyment, challenge) [30]. ...
... Overall, this study extends the ongoing discussion on algorithmic control [e.g., 50,82,91] by showing how its use within online gig-economy platforms can impact both worker well-being (in terms of their experience of threat and challenge technostressors) and behavior (in terms of their continuance intention and workaround use). It further answers emerging calls for research examining how algorithmic transparency relates to worker well-being [71,92]. From a practical perspective, our study provides valuable insights for organizations responsible for designing and implementing algorithms by highlighting the consequences that control design choices have on their workforce. ...
... We then provide an overview of the process of technostress, including the possible sources, appraisal, and outcomes. Similar to related studies [e.g., 16,71], our study uses Uber, the global ride-sharing company, as the contextualized example of a prominent gig economy platform that relies extensively on algorithmic control as part of its day-to-day operations [47,89]. ...
Article
Full-text available
This study examines how the use of algorithmic control within gig economy platforms relates to the well-being and behavior of workers. Specifically, we explore how two different forms of algorithmic control—gatekeeping and guiding—correspond with (positive) challenge technostressors and (negative) threat technostressors experienced by Uber drivers. We also examine the moderating impact of algorithmic control transparency on these relationships, as well as the outcomes of technostressors in terms of continuance intentions and workaround use. Based on a survey of 621 U.S.-based Uber drivers, we find that gatekeeping and guiding algorithmic control positively relate to both challenge and threat technostressors. The study bridges the literature on control and technostress by conceptualizing algorithmic control as a condition that puts workers under stress. This stress is found to contribute to important behavioral consequences pertaining to both continuance intentions and workaround use. Findings from our work suggest that gig economy organizations can use algorithmic control to enhance challenge technostressors for their workers, thereby contributing to the cultivation of a more committed workforce. Furthermore, we find evidence disputing the assumption that algorithmic control transparency can mitigate the negative effects of threat technostressors.
... Previous work has shown that algorithmic opacity may cause Uber drivers to experience uncertainties about financial compensation and ride assignments . Among some workers, such tensions trigger market-like responses, such as attempts to regain agency and work around the algorithms (e.g., gaming the system) (Cameron & Rahman, 2022;Möhlmann et al., 2021). ...
... Despite its manifold benefits for online labor platforms, algorithmic management is a doubleedged sword. Workers exposed to algorithmic management often report that they experience tensions in their work environment (Gal et al., 2020;Kellogg et al., 2020;Möhlmann et al., 2021;Page et al., 2017;Tilson et al., 2021;Wiener et al., 2021). For example, while gig workers often experience high levels of autonomy and flexibility (Rosenblat & Stark, 2016), they still feel surveilled and controlled through real-time surveillance (Newell & Marabelli, 2015;Zuboff, 2019). ...
... Previous research has reflected on how stakeholders outside organizations scrutinize algorithmic decision-making (Zuboff, 2019). Studies have also presented findings of platform workers' reactions to algorithmic management (Bucher et al., 2021;Cameron & Rahman, 2022;Curchod et al., 2020;Karanović et al., 2021;Möhlmann et al., 2021). For example, to feel part of a broader community and help one another navigate through the challenges imposed by their working environment, Uber drivers engage in informal communities (e.g., discussion in online forums and social media) (Lee et al., 2015;Möhlmann et al., 2021;Rosenblat & Stark, 2016). ...
Article
Full-text available
Algorithmic management may create work environment tensions that are detrimental to workplace well-being and productivity. One specific type of tension originates from the fact that algorithms often exhibit limited transparency and are perceived as highly opaque, which impedes workers' understanding of their inner workings. While algorithmic transparency may facilitate sensemaking, the algorithm's opaqueness may aggravate sensemaking. By conducting an empirical case study in the context of the Uber platform, we explore how platform workers make sense of the algorithms managing them. Drawing on Weick's enactment theory, we theorize a new form of sensemaking-algorithm sensemaking-and unpack its three sub-elements: (1) focused enactment, (2) selection modes, and (3) retention sources. The sophisticated, multi-step process of algorithm sensemaking allows platform workers to keep up with algorithmic instructions systematically. We add to previous literature by theorizing algorithm sensemaking as a mediator linking workers' perceptions about tensions in their work environment and their behavioral responses.
... Systems that carry out duties related to the organization of human work have been referred to as algorithmic management in the recent literature (Basukie, Wang, & Li, 2021;Cheng & Foley, 2019;Lee, Kusbit, Metsky, & Dabbish, 2015). Algorithmic tasks include, but are not limited to: (1) task allocation (Jarrahi & Sutherland, 2019); (2) pricingplatforms often use dynamic pricing, termed "primetime pricing" or "surge pricing" (Banerjee, Johari, & Riquelme, 2016;Chen, 2016); (3) control, enabled by peer-generated feedback/ratings and/or via tracking mechanisms (Duggan et al., 2020;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2021); (4) detection of fraud (Schildt, 2017); (5) evaluation (Duggan et al., 2020;Möhlmann et al., 2021); and (6) reward (Basili & Rossi, 2020;Prabowo, Sucahyo, Gandhi, & Ruldeviyani, 2019). ...
... Systems that carry out duties related to the organization of human work have been referred to as algorithmic management in the recent literature (Basukie, Wang, & Li, 2021;Cheng & Foley, 2019;Lee, Kusbit, Metsky, & Dabbish, 2015). Algorithmic tasks include, but are not limited to: (1) task allocation (Jarrahi & Sutherland, 2019); (2) pricingplatforms often use dynamic pricing, termed "primetime pricing" or "surge pricing" (Banerjee, Johari, & Riquelme, 2016;Chen, 2016); (3) control, enabled by peer-generated feedback/ratings and/or via tracking mechanisms (Duggan et al., 2020;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2021); (4) detection of fraud (Schildt, 2017); (5) evaluation (Duggan et al., 2020;Möhlmann et al., 2021); and (6) reward (Basili & Rossi, 2020;Prabowo, Sucahyo, Gandhi, & Ruldeviyani, 2019). ...
... In the case of positive learnings, trust in the platform and other peers (Chen et al., 2020;Wang, Asaad, et al., 2019) might increase commitment and effort (Basili & Rossi, 2020;Hofmann, Hartl, & Penz, 2017) and their attitudes towards the model will evolve in a positive direction. Conversely, unsuccessful interactions between PP and the platform might lead to loss of reciprocity, causing frustration (Jarrahi & Sutherland, 2019) and evoking agentic behaviours and feelings of gaming the system or cheating the algorithms (Jarrahi & Sutherland, 2019;Möhlmann et al., 2021) or voicing-up (Kougiannou & Mendonça, 2021). For example (Waldkirch, Bucher, Schou, & Grünwald, 2021) suggest that PP might pay surprisingly little attention to client satisfaction, looking for strategies of avoiding negative score (i.e. ...
Article
Peer (service) providers (PPs) are the frontline actors on sharing economy platforms (SEPs), and to date have received very limited attention in academic literature. These actors enter in economic exchanges with other stakeholders in the platform ecosystem by giving access to their underutilized assets – both tangible (vehicles, accommodation, clothes, etc.) and intangible (skills, time, etc.). Monetary compensation is the immediate and obvious benefit that these actors obtain. However, it has been shown that extrinsic benefits might not be sufficient to outweigh the evident costs associated with platform governance and spot-based transactions. This research is one of the first attempts to explore the satisfaction and loyalty of PPs, comparing capital and labour sharing platforms. We build on the social exchange theory and cost–benefit approach, developing the PP satisfaction model. Using structural equation modelling we analyse data from 205 PPs. We find that satisfaction with the platform and with the role of PP are related to loyalty. Sensitivity to the peer-to-peer business model is developed as a result of experiences and also impacts on loyalty.
... For instance, Deliveroo calculates delivery rates for drivers based on delivery performance such as time taken (Duggan et al., 2020;Woodcock, 2020). Ridesharing algorithms monitor drivers' actions (e.g., mobile phone use or sudden braking) during a ride, and issue alerts as needed (Möhlmann et al., 2021). They nudge drivers to continue working until a driver-specified earnings threshold is met. ...
... For example, using metrics to calculate one's performance is similar to a supervisor evaluating an employee's performance (Brynjolfsson & McAfee, 2014). Recent literature depicts algorithms as bosses (Möhlmann et al., 2021) or as prescriptive agents that can take on supervisory roles (Baird & Maruping, 2021). ...
... • Interactions occur through digital interfaces (e.g., an app) and are mediated by underlying algorithmic computations • Humans relate to algorithms as co-workers (Page et al., 2017) • Computational logic behind the algorithm's outputs is not always transparent (e.g., Bernstein, 2017;Dolata et al., 2021;Feuerriegel et al., 2020) Outputs of the algorithm • Perceived by the human as work instructions (e.g., pick up a passengers), work reward/ compensation (e.g., payment for a micro-task or a passenger ride), or nudges (e.g., warning for sudden braking) (Martin et al., 2014) • Human may not understand the work instructions and why the algorithm gives them (Pasquale, 2015) • Human may not be able to seek clarification or override the algorithm when they think the work instructions from the algorithm are inconsistent or incorrect (Marabelli et al., 2021) Role of the algorithm • Algorithms take on the role of an organisational member (Parent-Rocheleau & Parker, 2021); They are perceived by the human as a boss (Möhlmann et al., 2021) or supervisor (Baird & Maruping, 2021) • Human does not trust the algorithm (Glikson & Woolley, 2020) • Human may feel the algorithm is unfair (Parent-Rocheleau & Parker, 2021) and feel stifled and controlled by it (Gal et al., 2017;Zarsky, 2016) • Employees may resist algorithmic work (Kellogg et al., 2020) communication from the algorithm unclear to the human (Parent-Rocheleau & Parker, 2021;Rahman, 2021). Yet, the algorithm's work instructions may be mandatory and not following them might lead to penalties for the human as observed in the case of algorithmic ridesharing (Page et al., 2017). ...
Article
Full-text available
In algorithmic work, algorithms execute operational and management tasks such as work allocation, task tracking and performance evaluation. Humans and algorithms interact with one another to accomplish work so that the algorithm takes on the role of a co‐worker. Human–algorithm interactions are characterised by problematic issues such as absence of mutually co‐constructed dialogue, lack of transparency regarding how algorithmic outputs are generated, and difficulty of over‐riding algorithmic directive – conditions that create lack of clarity for the human worker. This article examines human–algorithm role interactions in algorithmic work. Drawing on the theoretical framing of organisational roles, we theorise on the algorithm as role sender and the human as the role taker. We explain how the algorithm is a multi‐role sender with entangled roles, while the human as role taker experiences algorithm‐driven role conflict and role ambiguity. Further, while the algorithm records all of the human's task actions, it is ignorant of the human's cognitive reactions – it undergoes what we conceptualise as ‘broken loop learning’. The empirical context of our study is algorithm‐driven taxi driving (in the United States) exemplified by companies such as Uber. We draw from data that include interviews with 15 Uber drivers, a netnographic study of 1700 discussion threads among Uber drivers from two popular online forums, and analysis of Uber's web pages. Implications for IS scholarship, practice and policy are discussed.
... Information systems (IS) research has mainly focused on the managerial operation of algorithms, for example, the design, implementation and use of algorithmic decision-making in organisations (Marabelli et al., 2021), and the coordination and control of the work process through computer algorithms (de Reuver et al., 2018;Hong et al., 2015;Möhlmann et al., 2020). Limited research has looked into broader social implications of platform work. ...
... While some emphasise workers' agency and even resistance to algorithmic management (Bucher et al., 2020;Chen, 2018;Jarrahi & Sutherland, 2019;Purcell & Brook, 2020), platforms and algorithms are often treated as objective, independent and technical entities that exert sovereign power over workers. Furthermore, algorithmic management renders managerial control less visible, hidden behind devices, which exert a combination of technical and normative control (Cram et al., 2020;Duggan et al., 2020;Kellogg et al., 2020;Möhlmann et al., 2020;Newlands, 2020) over workers by enacting 'information asymmetries' between the system and workers (Jarrahi & Sutherland, 2019;Rosenblat, 2016;Shapiro, 2018). ...
... In contrast, IS researchers tend to focus on transactional mechanisms or managerial processes on these platforms, in particular, the algorithmic matching, coordination and control (de Reuver et al., 2018;Hong et al., 2015;Möhlmann et al., 2020). Only a small number of IS papers interrogated the wider implications of algorithmic management beyond business benefits (Cram et al., 2020;Deng et al., 2016). ...
Article
Full-text available
Existing IS research on platform work has narrowly focused on the managerial operations of algorithmic management or its business implications. Limited research has paid attention to the scalar effects and societal implications of platform work. In this study, we address the phenomenon of ‘speed’ in the on‐demand economy through a qualitative study of Chinese food delivery workers. We construct a performative view of spatiotemporality to illustrate the reconfiguration of multiple spatiotemporal orders. The paper thus broadens the theorization of time and space in IS research and provides a more nuanced and critical understanding of platform work against the backdrop of structural inequality in platform capitalism.
... While a 'flesh-and-blood' supervisor is common sense in traditional work contexts, algorithms are used to control workers in the context of app work [3]. Such algorithmic control closely monitors and influences app workers' daily actions so that it aligns with the market logic goals of the app provider [4]. ...
... This study provides at least two contributions to the still infant literature on app work. First, we answer the call of several scholars [e.g., 4,16,17] to put forward theory on work design as a fruitful avenue to understand the impact of new technologies on work. In particular, we integrate algorithmic control with each of the three job characteristics that predict meaningful work according to job characteristics theory (hereafter JCT) [18]: skill variety (i.e., the extent to which an individual needs a diverse range of competencies and talents to do the job), task identity (i.e., the extent to which an individual can perform tasks within the job from start to finish with visible results), and task significance (i.e., the impact that the job has on other peoples' lives or work). ...
... To choose the preferred worker, algorithms that make a decision based on multiple parameters (e.g., distance to the restaurant) are used [22]. This daily decision-making process, however, is still a 'blackbox', as the food app providers are not transparent about the underlying logic of the algorithms' decisions [4]. ...
Conference Paper
Full-text available
App work disrupted our traditional understanding of work as it introduced new technologies, such as algorithmic control. Based on the job characteristics theory, we put forward an important drawback of algorithmic control and a practice that might mitigate it. We test whether algorithmic control obstructs experiences of meaningful work through a lack of motivating job characteristics and the buffering role of bottom-up work design (i.e. job crafting). We conduct a daily diary study among 51 Belgian food app workers and test within-person relationships. Results show that on days that app workers experience high algorithmic control, they perceive their work as less meaningful than on days with little algorithmic control. Although daily motivating job characteristics could not explain this negative relationship, we found job crafting to enable app workers in attaining motivating job characteristics and meaningful work. Thereby we emphasize the importance of both top-down and bottom-up work design in a strive for meaningful work
... In this paper, we take the perspective of blockchain governance (e.g., Beck et al. 2018;Chen et al. 2021;Lumineau et al. 2021;Murray et al. 2019;Ziolkowski et al. 2020b) and draw on the lens of governance via IT offered by the literature on digital platforms (e.g., Gregory et al. 2018;Moehlmann et al. 2021;Tiwana et al. 2010) which allows us to address the following research question: How are decentralized autonomous organizations governed via IT? For this purpose, we analyze five representative cases of DAOs (Aragon, Flare Networks, KyberDAO, MakerDAO, and MolochDAO) using techniques of grounded theory (Birks et al. 2013;Charmaz 2006;Gioia et al. 2013;Glaser and Strauss 1967;Walsh et al. 2015). ...
... Governance in these technology-enabled meta-organizations is characterized by balancing alignment desired by the platform owner with autonomy of users or complementors in the surrounding ecosystem (Gulati et al. 2012). Centralized platform owners and their software engineers often develop and maintain standardized boundary resources or algorithms to govern their community in the ecosystem (Ghazawneh and Henfridsson 2013;Moehlmann et al. 2021;Tiwana et al. 2010). Using algorithms to govern the behavior of individual actors and to ensure their alignment to the meta-organization's purpose can be summarized as governance via IT (Drnevich and Croson 2013;Gregory et al. 2018;Moehlmann et al. 2021). ...
... Centralized platform owners and their software engineers often develop and maintain standardized boundary resources or algorithms to govern their community in the ecosystem (Ghazawneh and Henfridsson 2013;Moehlmann et al. 2021;Tiwana et al. 2010). Using algorithms to govern the behavior of individual actors and to ensure their alignment to the meta-organization's purpose can be summarized as governance via IT (Drnevich and Croson 2013;Gregory et al. 2018;Moehlmann et al. 2021). ...
Conference Paper
Full-text available
A decentralized autonomous organization (DAO) is a distinct form of platform meta-organization that heavily relies on smart contracts running on blockchains to govern a distributed network of autonomous actors, thereby continuing the shift toward governance via IT. Motivated by the fact that this shift toward governance via IT in DAOs challenges established assumptions in the literature on IT governance, we explore how DAOs are governed via IT. For this purpose, we applied techniques of grounded theory to build inductive theory by analyzing five cases of DAOs (Aragon, Flare Networks, KyberDAO, MakerDAO, and MolochDAO) based on white papers, blog entries, and newspaper articles. Our findings implicate that DAOs governed via IT synthesize autonomy and alignment through the mechanism of "establishing algorithmic organization." At the same time, DAOs rely on a more pluralistic and decentralized form of algorithmic management through the mechanism of "taming algorithmic power."
... Prior research has established that work in many platform organizations is managed algorithmically for efficient coordination and control (e.g., Moehlmann et al. 2020), which is consistent with the perspective of agentic IS artifacts being delegated rights and responsibilities for decision-making and action (Baird and Maruping 2021). However, the shift from centralized to decentralized organizing manifested in DAOs that seek to empower Electronic copy available at: https://ssrn.com/abstract=3938489 ...
... In addition, our detailed findings about decentralized algorithmic management and decentralized management of algorithms in DAOs extend the Electronic copy available at: https://ssrn.com/abstract=3938489 literature on platform organizing (e.g., Beck et al. 2018;Gawer 2014;Gregory et al. 2018;Gulati et al. 2012;Moehlmann et al. 2020). ...
... Recent information systems (IS) research has advanced the key idea of IS artifacts as being agentic in nature, which is related to the ability of IS artifacts to accept rights and responsibilities for ambiguous tasks and outcomes under uncertainty and to decide and act autonomously (Baird and Maruping 2021). Baird and Maruping (2021) explain that agentic IS artifacts can now accept more complex and open-ended tasks related to sensing, cognition, and acting, and in fact recent in-depth empirical work on platform organizing through algorithmic management provides support for this claim (Moehlmann et al. 2020). While Baird and Maruping (2021) assume in their theorizing of IS delegation that one centralized human agent or agentic IS artifact is delegating tasks and responsibilities to another agent, they also discuss a possible challenge for this assumption that emerges in decentralized platform settings characterized by the agency of both human agents and agentic IS artifacts, in other words, a conjoined human-technology agency (Murray et al. 2020), that generates a collective power to achieve a common purpose. ...
... The use of AC is already commonplace with platform-based work in the gig economy (e.g., Duggan et al., 2020;Goldbach et al., 2018;Möhlmann et al., 2021). Enabling the effective "scaling of operations by [. . ...
... .] coordinating the activities of large, disaggregated workforces" (Mateescu & Nguyen, 2019, p. 3), AC represents a key ingredient to the success of gig economy platforms (Jabagi et al., 2019;Wood et al., 2019). A prime example is the gig firm Uber, which relies on an app-based AC system to control its remote workforce of some four million freelance drivers (e.g., Möhlmann et al., 2021;Scheiber, 2017). ...
... In light of these observations, prior research suggests that AC can be frustrating to gig workers, prompting them to question the legitimacy of AC practices, which can lead workers to discontinue working for a platform or violate controls (Curchod et al., 2020;Möhlmann et al., 2021) and cause serious harm to companies (e.g., Möhlmann & Henfridsson, 2019). As such, studying the notion of micro-level legitimacy judgements (Bitektine & Haack, 2015) broadly defined as an individual's assessment that "the actions of an entity [i.e., a gig company in the context of our study] are desirable, proper, or appropriate" (Suchman, 1995, p. 574) -may help link gig workers' perceptions of two predominant AC forms (gatekeeping and guiding) to their behavioural reactions (continuance intention and workaround use). ...
Article
Full-text available
Organisations increasingly rely on algorithms to exert automated managerial control over workers, referred to as algorithmic control (AC). The use of AC is already commonplace with platform-based work in the gig economy, where independent workers are paid for completing a given task (or “gig”). The combination of independent work alongside intensive managerial monitoring and guidance via AC raises questions about how gig workers perceive AC practices and judge their legitimacy, which could help explain critical worker behaviours such as turnover and non-compliance. Based on a three-dimensional conceptualisation of micro-level legitimacy tailored to the gig work context (autonomy, fairness, and privacy), we develop a research model that links workers’ perceptions of two predominant forms of AC (gatekeeping and guiding) to their legitimacy judgements and behavioural reactions. Using survey data from 621 Uber drivers, we find empirical support for the central role of micro-level legitimacy judgements in mediating the relationships between gig workers’ perceptions of different AC forms and their continuance intention and workaround use. Contrasting prior work, our study results show that workers do not perceive AC as a universally “bad thing” and that guiding AC is in fact positively related to micro-level legitimacy judgements. Theoretical and practical implications are discussed.
... The use of AC is already commonplace with platform-based work in the gig economy (e.g., Duggan et al., 2020;Goldbach et al., 2018;Möhlmann et al., 2021). Enabling the effective "scaling of operations by […] coordinating the activities of large, disaggregated workforces" (Mateescu & Nguyen, 2019, p. 3), AC represents a key ingredient to the success of gig economy platforms (Jabagi et al., 2019;Wood et al., 2019). ...
... Enabling the effective "scaling of operations by […] coordinating the activities of large, disaggregated workforces" (Mateescu & Nguyen, 2019, p. 3), AC represents a key ingredient to the success of gig economy platforms (Jabagi et al., 2019;Wood et al., 2019). A prime example is the gig firm Uber, which relies on an app-based AC system to control its remote workforce of some four million freelance drivers (e.g., Möhlmann et al., 2021;Scheiber, 2017). ...
... In light of these observations, prior research suggests that AC can be frustrating to gig workers, prompting them to question the legitimacy of AC practices, which can lead workers to discontinue working for a platform or violate controls (Curchod et al., 2020;Möhlmann et al., 2021) and cause serious harm to companies (e.g., Möhlmann & Henfridsson, 2019). As such, studying the notion of micro-level legitimacy judgements (Bitektine & Haack, 2015)-broadly defined as an individual's assessment that "the actions of an entity [i.e., a gig company in the context of our study] are desirable, proper, or appropriate" (Suchman, 1995, p. 574)-may help link gig workers' perceptions of two predominant AC forms (gatekeeping and guiding) to their behavioral reactions (continuance intention and workaround use). ...
Preprint
Full-text available
Organizations increasingly rely on algorithms to exert automated managerial control over workers, referred to as algorithmic control (AC). The use of AC is already commonplace with platform-based work in the gig economy, where independent workers are paid for completing a given task (or ‘gig’). The combination of independent work alongside intensive managerial monitoring and guidance via AC raises questions about how gig workers perceive AC practices and judge their legitimacy, which could help explain critical worker behaviors such as turnover and non-compliance. Based on a three-dimensional conceptualization of micro-level legitimacy tailored to the gig work context (autonomy, fairness, and privacy), we develop a research model that links workers’ perceptions of two predominant forms of AC (gatekeeping and guiding) to their legitimacy judgments and behavioral reactions. Using survey data from 621 Uber drivers, we find empirical support for the central role of micro-level legitimacy judgements in mediating the relationships between gig workers’ perceptions of different AC forms and their continuance intention and workaround use. Contrasting prior work, our study results show that workers do not perceive AC as a universally ‘bad thing’ and that guiding AC is in fact positively related to micro-level legitimacy judgements. Theoretical and practical implications are discussed.
... AM can be conceived as the deployment of algorithms for making and executing decisions affecting labor (Duggan et al., 2020). Growing scholarly interest in AM on DLPs yields related literature on the topic in multiple disciplines, such as information systems (e.g., Cram et al., 2022;Möhlmann et al., 2021), computer science (e.g., Lee et al., 2015;Yu et al., 2017), human resource management (e.g., Duggan et al., 2020;Meijerink et al., 2021), and organizational theory (e.g., Kellogg et al., 2020;Kinder et al., 2019). AM has also been studied under related terms, such as algorithmic control (e.g., Kellogg et al., 2020), technologymediated control (e.g., Cram & Wiener, 2020), algorithmic governance (e.g., Bucher et al., 2021), or people analytics (e.g., Gal et al., 2017). ...
... There seems to be no agreement on specific instantiations of AM, thereby not clearly distinguishing between AM functions, mechanisms, practices, and features. For instance, Möhlmann et al. (2021) specify AM functions as coordination (i.e., algorithmic matching), and control (i.e., algorithmic control), whereas Lee et al. (2015) specify AM features, such as work assignment (i.e., driver-passenger assignment algorithms), informational support (i.e., dynamic in-app display of surge-priced areas), and performance evaluation (i.e., rating systems and acceptance rates that track driver performance). Without attempting to preempt the analysis, we note here, that the multiplicity of AM instantiations is in line with our participants' perceptions of AM, which manifest in forms of "the system", "the platform", "mechanisms", "functions", "applications", "processes" and "features". ...
Conference Paper
Full-text available
On digital labor platforms, interactions between workers and clients are algorithmically managed. Previous research found that algorithmic management can disadvantage workers. In this paper, we empirically examine algorithmic unfairness from a sociotechnical perspective. Specifically, we conduct online focus groups with 23 workers who directly interact with algorithmic management practices on digital labor platforms. In using grounded theory methodology, we pursue to understand how algorithmic management promotes unfairness on digital labor platforms. Our emergent theory understands algorithmic unfairness as algorithmic management practices that give rise to systematic disadvantages for workers. Algorithmic management practices either automate decisions or automate the delegation of decisions. Workers experience systematic disadvantages in the form of devaluation, restriction, and exclusion. Our findings serve as a starting point for mitigating algorithmic unfairness in the future.
... For instance, Upwork provides a desktop app that can automatically take screenshots and track workers' activities on hourly projects. 5 From the perspective of employers, monitoring allows them to track workers' progress and intervene when necessary (Möhlmann et al. 2021). Prior studies show that monitoring can improve workers' productivity (Hubbard 2000, Duflo et al. 2012. ...
... When the monitoring policy is opaque, workers are uncertain about what information is collected. Accordingly, they may be skeptical about whether the information recorded by the monitoring system can be used as evidence of their work, thereby leading to uncertainty of their compensation (Möhlmann et al. 2021). Conversely, a transparent monitoring policy provides workers the necessary information to evaluate whether the recorded information can protect them from payment disputes (Karwatzki et al. 2017). ...
Article
Monitoring is ubiquitous in the gig economy wherein the workforce is geographically dispersed. However, workers are often reluctant to be monitored because of privacy concerns, resulting in a hidden economic cost for employers as workers demand higher wages for monitored jobs. We investigate how three common dimensions of monitoring affect workers’ willingness to accept monitored jobs through online experiments on two gig economy platforms. The three dimensions of monitoring are intensity (how much information is collected), transparency (whether the monitoring policy is disclosed to workers), and control (whether workers can remove sensitive information). We find that, as the monitoring intensity increases, workers become less willing to accept monitoring because of elevated privacy concerns. Furthermore, being transparent about the monitoring policy increases workers’ willingness to accept monitoring only when the monitoring intensity is low. Interestingly, providing control over high-intensity monitoring does not significantly reduce workers’ privacy concerns, rendering this well-intentioned policy ineffective. Finally, females are more willing to accept monitored jobs than males as they perceive higher payment protection from monitoring and have lower privacy concerns. On average, the hourly wage compensation required for gig workers to accept monitoring is 1.6∼1.8 dollars, which amounts to roughly 28.6%∼37.5% of their average hourly wage.
... Thus, enhancing a platform's worker retention and reducing platform turnover is a central challenge that needs in-depth investigation (Williams et al. 2021). While research has started to shed light on the lived experience of workers, a large share of studies still focuses on negative experiences (Caza et al. 2021;Möhlmann et al. 2020;Petriglieri et al. 2019); however, insight on how to retain workers is still underdeveloped. ...
... With a view to for future projects, clients can also give workers feedback on their performance (Durward et al. 2020;Strunk et al. 2021). Several studies support the relevance of transparency and feedback for workers' positive lived experienced (Durward et al. 2020;Möhlmann et al. 2020;Parent-Rocheleau and Parker 2021). In contrast, central challenges are no client feedback nor process transparency, which often lead to frustration and feelings of marginalization Fieseler et al. 2019;Parent-Rocheleau and Parker 2021;Strunk et al. 2021). ...
Conference Paper
This study investigates how supportive social relationships experienced by workers on online working platforms counteract their platform turnover. Keeping workers active and reducing turnover is a key challenge in platform work. Still, to date our knowledge on retaining workers is limited. Organizational literature concludes that supportive organizational environments are strong predictors of reduced turnover. Although platform work lacks organizational social environments, studies shed light on the social context digital platforms do provide within their platform ecosystems. Building on social identity and social exchange theory, I investigate how supportive relationships experienced within this ecosystem reduce platform turnover. Drawing on a two-wave survey with 652 workers, I found that the platform ecosystem’s social support (particularly from the platform provider and from virtual peer communities) shapes workers’ affective commitment to the platform, which reduces turnover intentions. From an ecosystem perspective these findings make several contributions to the social dynamics in platform work.
... For example, we can think of an Uber and the algorithmic management it uses to nudge its drivers [62]. Uber uses nudging as an element of algorithmic control -'the use of algorithms to monitor platform workers' behavior and ensure its alignment with the platform organization's goals' to offer high-quality service to customers [62:18]. ...
... Building on the insights from these data, Uber's algorithm learns when exactly to nudge drivers and iteratively improve its attempts to influence drivers' behavior. Before introduction of these techniques, drivers complained about too many or even nagging attempts to nudge them [62]. Simple interface-based techniques did not always work. ...
... Amongst other purposes, this involves the use of software algorithms that decide which gig workers are allowed access to the online marketplace of a platform firm (Jarrahi & Sutherland, 2019;Veen et al., 2019), those that compile and publish reputation information from client ratings and deactivate those workers who drop below standards (Rosenblat & Stark, 2016), and thoses that determine and adjust pay rates for gig workers and calculate surge prices (Rosenblat, 2019). Together these processes are referred to as algorithmic management, a system of control that relies on machine-readable data and algorithms to automate HR-related decision-making (Duggan et al., 2019;Lee et al., 2015a;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2020). Algorithmic management comes in many different shapes and forms, and may range from the use of relatively simple sorting algorithms up to self-learning algorithms that fully automate decision-making and execution without (much) human involvement (Cheng & Hackett, 2021;Strohmeier & Piazza, 2016). ...
... At the time of this writing, the COVID-19 outbreak was attracting renewed attention to the status and outcomes of gig workers and to the power of digital labor platforms (Hasija et al., 2020;Moulds, 2020). But for many years the digital labor platform phenomenon generated far more interest from researchers in information systems (Jarrahi & Sutherland, 2019;Lee et al., 2015a;Möhlmann et al., 2020;Sutherland & Jarrahi, 2018), economics (Abraham et al., 2018;Dube et al., 2020;Horton, 2017;Pallais, 2014), and sociology (Gandini, 2019;Rosenblat, 2019;Shapiro, 2018;Schor et al., 2019;Vallas & Schor, 2020;Van Doorn, 2020;Veen et al., 2019;Wood et al., 2019a;Wood et al., 2019b) than from scholars trained in human resource management and industrial/organizational psychology (Ashford et al., 2018;Duggan et al., 2020;Kuhn, 2016;Meijerink & Keegan, 2019). Much of this literature has been concerned with the precarity and vulnerability of gig workers, whereas others have envisioned platforms as "marketplaces" of rational actors and as ideal test cases for economic theory (Rietveld & Schilling, 2021). ...
Chapter
Full-text available
This work examines the intersection between traditional human resource management and the novel employment arrangements of the expanding gig economy. While there is a substantial multidisciplinary literature on the digital platform labor phenomenon, it has been largely centered on the experiences of gig workers. As digital labor platforms continue to grow and specialize, more managers, executives, and human resource practitioners will need to make decisions about whether and how to utilize gig workers. Here the authors explore and interrogate the unique features of human resource management (HRM) activities in the context of digital labor platforms. The authors discuss challenges and opportunities regarding (1) HRM in organizations that outsource labor needs to external labor platforms, (2) HRM functions within digital labor platform firms, and (3) HRM policies and practices for organizations that develop their own spin-off digital labor platform. To foster a more nuanced understanding of work in the gig economy, the authors identify common themes across these contexts, highlight knowledge gaps, offer recommendations for future research, and outline pathways for collecting empirical data on HRM in the gig economy.
... Amongst other purposes, this involves the use of software algorithms that decide which gig workers are allowed access to the online marketplace of a platform firm (Jarrahi & Sutherland, 2019;Veen et al., 2019), those that compile and publish reputation information from client ratings and deactivate those workers who drop below standards (Rosenblat & Stark, 2016), and thoses that determine and adjust pay rates for gig workers and calculate surge prices (Rosenblat, 2019). Together these processes are referred to as algorithmic management, a system of control that relies on machine-readable data and algorithms to automate HR-related decision-making (Duggan et al., 2019;Lee et al., 2015a;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2020). Algorithmic management comes in many different shapes and forms, and may range from the use of relatively simple sorting algorithms up to self-learning algorithms that fully automate decision-making and execution without (much) human involvement (Cheng & Hackett, 2021;Strohmeier & Piazza, 2016). ...
... At the time of this writing, the COVID-19 outbreak was attracting renewed attention to the status and outcomes of gig workers and to the power of digital labor platforms (Hasija et al., 2020;Moulds, 2020). But for many years the digital labor platform phenomenon generated far more interest from researchers in information systems (Jarrahi & Sutherland, 2019;Lee et al., 2015a;Möhlmann et al., 2020;Sutherland & Jarrahi, 2018), economics (Abraham et al., 2018;Dube et al., 2020;Horton, 2017;Pallais, 2014), and sociology (Gandini, 2019;Rosenblat, 2019;Shapiro, 2018;Schor et al., 2019;Vallas & Schor, 2020;Van Doorn, 2020;Veen et al., 2019;Wood et al., 2019a;Wood et al., 2019b) than from scholars trained in human resource management and industrial/organizational psychology (Ashford et al., 2018;Duggan et al., 2020;Kuhn, 2016;Meijerink & Keegan, 2019). Much of this literature has been concerned with the precarity and vulnerability of gig workers, whereas others have envisioned platforms as "marketplaces" of rational actors and as ideal test cases for economic theory (Rietveld & Schilling, 2021). ...
Chapter
Full-text available
This work examines the intersection between traditional human resource management and the novel employment arrangements of the expanding gig economy. While there is a substantial multidisciplinary literature on the digital platform labor phenomenon, it has been largely centered on the experiences of gig workers. As digital labor platforms continue to grow and specialize, more managers, executives, and human resource practitioners will need to make decisions about whether and how to utilize gig workers. Here we explore and interrogate the unique features of human resource management (HRM) activities in the context of digital labor platforms. We discuss challenges and opportunities regarding 1) HRM in organizations that outsource labor needs to external labor platforms, 2) HRM functions within digital labor platform firms, and 3) HRM policies and practices for organizations that develop their own spin-off digital labor platform. To foster a more nuanced understanding of work in the gig economy, we identify common themes across these contexts, highlight knowledge gaps, offer recommendations for future research, and outline pathways for collecting empirical data on HRM in the gig economy.
... Deviating from its prescriptions has consequences that include a lower rating, loss of income and ultimately being laid off (Page et al. 2017). Yet, the long-term use of these algorithms has demonstrated that drivers cannot sever customers effectively if they follow the (rigid) algorithm blindly (Möhlmann et al. 2020). Clearly the Uber algorithm wasn't designed or rolled out to upset customers (or make drivers uncomfortable) but its use in practice unveiled these issues. ...
... We have the knowledge and business-related background on how the workforce should (or should not) be managed. This would also add to (and build on) the growing (IS) body of literature on algorithms at work, for instance in Uber contexts (Möhlmann et al. 2020;Page et al. 2017). This would help addressing questions such as: how is it possible to incorporate human vetting of ADMS in novel and emergent surveillance settings? ...
Article
Full-text available
in this viewpoint article we discuss algorithmic decision-making systems (ADMS), which we view as organizational sociotechnical systems with their use in practice having consequences within and beyond organizational boundaries. We build a framework that revolves around the ADMS lifecycle and propose that each phase challenges organizations with "choices" related to technical and processual characteristics-ways to design, implement and use these systems in practice. We argue that it is important that organizations make these strategic choices with awareness and responsibly, as ADMS' consequences affect a broad array of stakeholders (the workforce, suppliers, customers and society at-large) and involve ethical considerations. With this article we make two main contributions. First, we identify key choices associated with the design, implementation and use in practice of ADMS in organizations, that build on past literature and are tied to timely industry-related examples. Second, we provide IS scholars with a broad research agenda that will promote the generation of new knowledge and original theorizing within the domain of the strategic applications of ADMS in organizations.
... As our variables of politeness density, authentication, sales amount, review rating, tenure, year, prior experience, and amount of information were power-law distributed, we employed a natural logarithm transformation [3,31]. The correlation coefficients among these variables were less than 0.7; the VIF values for all variables were less than 3 (range from 1.01 to 1.15), with an average value of 1.07, thus, multicollinearity did not seem to be an issue in this paper [32,33]. Because our dependent variable was a binary variable, according to prior research, such as Lv et al. [4] and Coussement et al. [34], we used the binary logistics regression model to analyze the data collected from ZBJ. ...
Article
Full-text available
This study examined the effect of politeness, as a key reflection of linguistic features of conversation in the online labor marketplace, on hiring behavior. Drawing on the politeness theory, a non-linear relationship was theorized. A hypothesis was put forward and examined against a large-scale archival dataset from a Chinese online labor market. Using an econometric model, the results demonstrated that there was an inverted U-shaped relationship between politeness and hiring decisions. The study offers theoretical implications to the online labor market literature and politeness theory by providing empirical insights on the role of politeness in hiring decision. In addition, our findings offer beneficial and practical contributions for vendors and platform operators.
... This fuels competition and spirals wages downwards. Besides the intense competition and limited opportunities for signalling work quality, which we describe in this study quantitatively, the literature has discussed the role of algorithmic control [27,83] and the organisation mechanisms of work in the platform economy [30,84] as drivers of (adverse) outcomes for remote platform workers. ...
Article
Full-text available
The Covid-19 pandemic has led to the rise of digitally enabled remote work with consequences for the global division of labour. Remote work could connect labour markets, but it might also increase spatial polarisation. However, our understanding of the geographies of remote work is limited. Specifically, in how far could remote work connect employers and workers in different countries? Does it bring jobs to rural areas because of lower living costs, or does it concentrate in large cities? And how do skill requirements affect competition for employment and wages? We use data from a fully remote labour market—an online labour platform—to show that remote platform work is polarised along three dimensions. First, countries are globally divided: North American, European, and South Asian remote platform workers attract most jobs, while many Global South countries participate only marginally. Secondly, remote jobs are pulled to large cities; rural areas fall behind. Thirdly, remote work is polarised along the skill axis: workers with in-demand skills attract profitable jobs, while others face intense competition and obtain low wages. The findings suggest that agglomerative forces linked to the unequal spatial distribution of skills, human capital, and opportunities shape the global geography of remote work. These forces pull remote work to places with institutions that foster specialisation and complex economic activities, i. e. metropolitan areas focused on information and communication technologies. Locations without access to these enabling institutions—in many cases, rural areas—fall behind. To make remote work an effective tool for economic and rural development, it would need to be complemented by local skill-building, infrastructure investment, and labour market programmes.
... One vivid example of PSD that touches ethical boundaries is the case of Uber and its algorithmic management of workers (Möhlmann, Zalmanson, Henfridsson, & Gregory, 2021). Because of the nature of the business relationship with its drivers (i.e., independent contractors), Uber is not as strongly tied by employment laws (Scheiber, 2017). ...
Article
Persuasive System Design (PSD) is an umbrella term for designs in information systems (IS) that can influence the attitude, behavior, or decision-making of people for better or worse. On the one hand, PSD can improve the engagement and motivation of users to change their attitude, behavior, or decision-making way favorably and thus help them achieve the desired outcome, thus improving the users’ well-being. On the other hand, PSD can also be misused and lead to unethical and undesirable outcomes such as disclosing unnecessary information or agreeing to terms that are unfavorable for the user, which in return can negatively impact the users’ well-being. These powerful persuasive designs can involve concepts such as gamification, gamblification, or digital nudging, which all have become well-renowned during the last years and have been implemented successfully across different sectors including education, e-health, e-governance, e-finance, or digital privacy contexts. However, such persuasive influence on individuals raises ethical questions as PSD can impair the autonomy of users or persuade the user towards the goals of a third party and hence lead to unethical decision-making processes and outcomes. In human-computer interaction, this is especially significant with the advent of advanced artificial intelligence. These novel technologies allow to influence the decision-making of users, gather data, and profile users as well as persuade them towards unethical outcomes. These unethical outcomes can lead to psychological and emotional damage in users. To understand the role of ethics in persuasive system design, we conduct an exhaustive systematic literature analysis and a set of 20 interviews to compile an overview of ethical considerations for persuasive system design. Furthermore, we then derive potential propositions for more ethical PSD and shed light on potential research gaps.
... For example, some gig workers prefer to play 'relational games', i.e., they spend efforts in building relationships with clients; others prefer to play 'efficiency' games, i.e., they set boundaries with clients and focus on maximizing how much they can be paid (Cameron, 2022). Some gig workers display typical behaviours of self-employed workers, e.g., resisting platform's instructions and algorithmic control or switching between alternative platforms; others enact organization-like responses, e.g., expressing attachment, gratitude, and loyalty towards the platform (Möhlmann et al., 2021). ...
Article
Full-text available
On online labour platforms, algorithmic scores are used as indicators of freelancers' work quality and future performance. Recent studies underscore that, to achieve good scores and secure their presence on platforms, freelancers respond to algorithmic control in different ways. However, we argue, to fully understand how freelancers deal with algorithmic scores, we first need to investigate how they interpret scores and, more specifically, what scores can do for them, i.e., perceived algorithmic affordances and constraints. Our interviews and other qualitative data collected with knowledge intensive gig workers on a major platform allow us to explain how the perceived affordances of algorithms (i.e., barrier, individual visibility, self-extension, rule of the game) act as mechanisms that explain different behavioural and emotional responses over time. Our work contributes to the current debate on the positive and negative consequences of algorithmic work by portraying the fundamental role paid by the individual interpretation of algorithmic scores and by integrating the affordance perspective into our understanding of algorithmic work.
... It was not always the case as other interviewees expressed lack of faith in the customer service to make fair judgement, which will be elaborated later. For gig workers who are app-based and have little or no access to human managers, customer service is an important or even the single source of organizational support (Möhlmann et al., 2021). Perceived organizational support could moderate the effects of job demands on health and performance related outcomes (Cheng and Chen, 2020). ...
Preprint
Full-text available
The COVID-19 pandemic has given the global e-commerce market a strong boost, of which China has the largest share and is growing rapidly. Concerns have been raised about intensified work stress and its consequences on health and safety among Chinese couriers. Sociological research of work and occupations has offered important insights into the labour process and politics of the gig economy, although how exactly the workers perceive and respond to technology-driven structural changes remains less clear. We conducted 14 semi-structured interviews with frontline couriers in May-June 2021 in China and interpreted the emerged themes following the Job Demands-Resources (JD-R) model. Four major work-stressor themes were identified: customer sovereignty, algorithmic management, economic precarity and networked support. These work conditions rarely worked alone. Technological, managerial and customer controlling mechanisms reinforced each other and increased work stress. In the absence of adequate organizational support, workers found support and resources through personal networks.
... The available data collected in one stage can be used to influence algorithms that are implemented by platform owners to make decisions in other stages. A prominent example is the use of workers' quality indicators gained in the quality control stage to determine the available work opportunities in the matchmaking stage (Möhlmann et al. forthcoming;Rani et al. 2021). ...
Conference Paper
Full-text available
Digital labor platforms (DLPs) enable new work arrangements by connecting workers who sell their labor, with clients, who purchase their services through the platform. As platform governance in the multi-actor setting is complex, we provide platform actors, policy makers, and researchers with an overview of the possible configurations of platform governance on DLPs. By focusing on "who" makes decisions, the influence platform owners, worker, and clients have on the platform is surfaced. Our taxonomy differentiates decision rights throughout the service transaction process, which includes matchmaking, price setting, scope setting, process control direct exchange and quality control. We classify 106 DLPs and find that platform owners can exert considerable influence or, alternatively, empower workers and clients by granting them decisions rights. A cluster analysis results in three archetypes which show that the influence of the platform owner increases from open marketplace DLPs, to cooperative DLPs, to platform-controlled DLPs.
... Thus far, organizations have installed AI supervisors mostly in middle management positions in which they translate orders or goals from upper management into daily instructions for lower-level employees (Wesche & Sonderegger, 2019). More recently, however, organizations have started to give AI supervisors more sovereign authority, and gig economy companies such as Uber now commonly rely on AI to instruct their workers and even punish them for alleged misbehavior (Möhlmann, Zalmanson, Henfridsson, & Gregory, 2021). ...
... Another major source of stress frequently mentioned in the literature is intensive algorithmic control imposed by digital platforms. While the platform workers are usually classified as independent contractors, the mechanism for managing this 'decentralized' workforce is very much centralized (de Vaujany et al., 2021), except that the algorithms employed by the platforms are perceived as the 'boss' (Möhlmannn et al., 2021). Under the omniscient 'algorithmic gaze' (Newlands, 2020), platform workers are subject to various motivating incentives and disciplinary sanctions. ...
Article
Full-text available
This study draws upon organizational psychology and platform labor research to investigate how socio‐psychological factors affect the mental well‐being of platform workers and help them cope with the challenges of work. Based on a survey study of 500 food‐delivery workers (‘riders’) in China, we provide quantitative evidence of workers' ambivalent subjective experiences that complements the predominantly qualitative account in the extant literature. In particular, we assess the complex relationships between meaningfulness of work, autonomy at work, self‐perceived competence, and workers' subjective well‐being. Our data also show that the stress‐buffering effect of social support mainly comes from the riders' familial contact and their online group chat with other workers. Overall, despite the well‐documented precarity and stress in platform work, the riders in our sample appear to be able to mobilize inner and relational resources to achieve a relatively high‐level mental well‐being.
... Besides these tangible AS that link the physical world to the information world (Barrett 2006), we note a growing number of intangible AS in the form of software systems that operate either entirely in the background or at the interface with humans. Examples are intelligent chatbots, smart contracts, and recommender systems (Murray et al. 2021a;Pfeiffer et al. 2020;Rutschi and Dibbern 2020;Wang et al. 2019a, b), as well as algorithmic management and control systems, such as the ones used by Uber and other gig economy firms to manage their digital workforce (Cram and Wiener 2020; Möhlmann et al. 2021;Wiener et al. 2021). ...
... The tension is further reinforced by virtue of a superapp being both a market that matches (value is created from optimizing interactions between demand and supply) and a firm that controls (value is created by coordinating constitutive human actors and non-human actants towards organizational goals) (Möhlmann et al., 2021), and is most salient vis-à-vis the perceived contradiction arising from the imposition of a variety of human-resource management like activities that are being algorithmically determined absent a clearly recognisable employment relationship in the traditional sense (Meijerink & Keegan, 2019). In attempting to formalize the informal economy of independent contract labour without extending the benefits and protection of traditional employment, superapps are more likely to (inadvertently) further dilute the already vulnerable's lack of autonomy and agency (Sasikumar & Sersia, 2021). ...
Thesis
Full-text available
On account of superapps' novelty (as both a product and a form of organizing value), and the global impact of emblematic superapps such as WeChat and Alipay, this paper seeks to understand the superapp phenomenon as a meta organizing principle of value, and how its innovation trajectories may be curated such that its value architecting processes contribute to both private profit and societal well-being. Via an integrative literature review, this research proposes three conceptual linkages (complexity, interactivity, responsibility) that paradigmatically reframes the relationship between the superapp phenomenon, the corporate social innovation construct, and the value (co-)creation framework. Such a reframing affords (i) academics phenomenologically grounded entry points to more thoughtfully investigate various aspects of the superapp; (ii) practitioners tool to better facilitate more responsible means of harvesting the generative tensions inhered in the networked relationalities embedded within the superapp's ecosystem; and (iii) policy makers guidance vis-à-vis the creation of regulations that nurtures superapps' potential for the greater good.
... The increasing availability of data and sophistication of algorithms (including the rebirth of machine learning / neural networks [19]) has enabled more uses and misuses of algorithmically controlled, automated decision-making (ADM, for short) [15]. The scaling of such innovations has happened in a context where algorithms, which still bear little accountability [8], [20], are involved in automating news recommendations [21], [22], advertising [23], but also work [24], [25] [26] and other highly sensitive processes such as social ranking, crime prediction, and bail, parole and criminal sentencing [27]- [29] . ...
... In the retail sector scheduling algorithms seek to minimise retailer labour use while also minimising the numbers of workers who work enough hours to qualify for the enhanced benefits, resulting in instability of hours and income (Schulte, 2020;Ton, 2012). Algorithmic management tools with similar effects are being developed and deployed in the nascent 'gig work' sector of technologically mediated, hyper flexible employment (Duggan et al., 2020;Möhlmann et al., 2021). ...
Article
Full-text available
Artificial intelligence (AI) is widely heralded as a new and revolutionary technology that will transform the world of work. While the impact of AI on human resource (HR) and people management is difficult to predict, the article considers potential scenarios for how AI will affect our field. We argue that although popular accounts of AI stress the risks of bias and unfairness, these problems are eminently solvable. However, the way that the AI industry is currently constituted and wider trends in the use of technology for organising work mean that there is a significant risk that AI use will degrade the quality of work. Viewing different scenarios through a paradox lens, we argue that both positive and negative visions of the future are likely to coexist. The HR profession has a degree of agency to shape the future if it chooses to use it; HR professionals need to develop the skills to ensure that ethics and fairness are at the centre of AI development for HR and people management.
... Even in the final stage when workers file disputes to try to salvage their ratings, the platform must dedicate resources to arbitrating these disputes. Studies find that updating algorithms to try to counteract workers' covert resistance takes considerable time, money, talent, and energy (Gillespie 2018, Keller 2018, Möhlmann et al. 2020). Thus, our study suggests that platforms bear increased costs when workers devise and implement covert resistance tactics. ...
Article
Full-text available
Existing literature examines control and resistance in the context of service organizations that rely on both managers and customers to control workers during the execution of work. Digital platform companies, however, eschew managers in favor of algorithmically mediated customer control—that is, customers rate workers, and algorithms tally and track these ratings to control workers’ future platform-based opportunities. How has this shift in the distribution of control among platforms, customers, and workers affected the relationship between control and resistance? Drawing on workers’ experiences from a comparative ethnography of two of the largest platform companies, we find that platform use of algorithmically mediated customer control has expanded the service encounter such that organizational control and workers’ resistance extend well beyond the execution of work. We find that workers have the most latitude to deploy resistance early in the labor process but must adjust their resistance tactics because their ability to resist decreases in each subsequent stage of the labor process. Our paper, thus, develops understanding of resistance by examining the relationship between control and resistance before, during, and after a task, providing insight into how control and resistance function in the gig economy. We also demonstrate the limitations of platforms’ reliance on algorithmically mediated customer control by illuminating how workers’ everyday interactions with customers can influence and manipulate algorithms in ways that platforms cannot always observe.
... Technology studies have addressed how technologies evoke emotions such as stress and anxiety (Barley et al. 2011;Bailey et al. 2012;Hinds and Bailey 2003). For example, employees (e.g., at Uber and eBay) may experience social isolation arising from being controlled and exploited by the decisions made by "algorithm bosses" (Curchod et al. 2020;Möhlmann et al. 2020). However, they have tended to overlook the moral emotions that can arise when technologies infringe the values of work (Evans 2021). ...
Article
Full-text available
Technologies are known to alter social structures in the workplace, reconfigure roles and relationships, and disrupt status hierarchies. However, less attention has been given to how an emerging technology disrupts the meaning and moral values that tether people to their work and render it meaningful. To understand how workers respond to such an emerging technology, we undertook an inductive, qualitative study of military personnel working in unmanned aerial vehicles, or drone operations, for the U.S. Air Force. We draw on multiple data sources, including personal diaries kept by personnel involved in drone operations. We identified three characteristics of drone technology: 'remote-split' operations, remote piloting of unmanned vehicles, and interaction through iconic representations. Our analysis suggests that drone technology has revolutionized warfare by 1) creating distanciated intimacy, 2) dissolving traditional spatio-temporal boundaries between work and personal life, and 3) redefining the legal and moral parameters of work. Drone program workers identified with these changes to their working environment in contradictory ways, which evoked emotional ambivalence about right and wrong. However, their organization gave them little help in alleviating their conflicting feelings. We illuminate how workers cope with such ambivalence when a technology transforms the meaning and morality of their work. We extend theory by showing that workers' responses to a changed working environment as a result of a remote technology are not just based on how the technology changes workers' tasks, roles and status, but also on how it affects their moral values.
... Crowd work platforms attract a heterogenous group of workers often relying on platforms as additional income source, with a majority of workers holding academic degrees but reporting low income (Berg, Furrer, Harmon, Rani, & Silberman, 2018;Deng, Joshi, & Galliers, 2016). With recent reports indicating that a significant share of the working population in various countries already participated in a type of crowd work (Berg et al., 2018;Huws et al., 2016) and the number of platforms and workers steadily rising all over the world (Durward et al., 2020;Kässi & Lehdonvirta, 2018) online labor platforms increasingly shape the future of work (Brynjolfsson & Mitchell, 2017;Möhlmann, Zalmanson, Henfridsson, & Gregory, 2021). ...
Article
This paper identifies antecedents to frustration arising in micro-task crowd work (MTCW) and measures their impact on a frustrating work experience. Surprisingly, research to date has said little on frustration in MTCW, despite it being a central challenge to working persistently in these increasingly important socio-technical work environments. Guided by an experiential computing perspective, we applied a mixed-method approach to reveal how the interplay between platforms and other workers’ behavior shapes frustration experienced in MTCW. We present qualitative and quantitative findings on frustration in virtual MTCW. Based on 96 narratives, we qualitatively identify three dominant antecedents to frustration: technical problems in conducting tasks, platform competition and, the inability to disconnect from work. We subsequently conducted a quantitative two wave survey (N=181), to test the factors’ impacts on MTCW frustration. The results confirm the three factors and additionally reveal that autonomy largely reduces frustration. However, autonomy also reinforced frustration arising from the inability to disconnect from work. Thus, workers’ inability to disconnect from work can drive excessive availability that lures them into taking on additional tasks. Thereby, workers experience additional frustration and underutilize the benefits of autonomy in crowd work. Our manuscript offers three central contributions: First, we identify frustration as pivotal in the experience of MTCW and its antecedents. Second, we shed light on the heterogenous role of autonomy in relation to MTCW frustration. Finally, we link the findings on frustrating MTCW experience to experiential computing.
... Technology studies have addressed how technologies evoke emotions such as stress and anxiety (Barley et al. 2011;Bailey et al. 2012;Hinds and Bailey 2003). For example, employees (e.g., at Uber and eBay) may experience social isolation arising from being controlled and exploited by the decisions made by "algorithm bosses" (Curchod et al. 2020;Möhlmann et al. 2020). However, they have tended to overlook the moral emotions that can arise when technologies infringe the values of work (Evans 2021). ...
Article
Full-text available
Digital platform technology enables circular business models that facilitate the reduction, reuse, and recycling of resources and materials across large ecosystems of platform actors. However, little is currently known about the inner workings of such platforms and how they are organized. Framing these platforms as meta-organizations, this study examines the orchestration mechanisms deployed by platform owners to facilitate economic value creation with a circular business model among a large group of actors. Building on an inductive analysis of 10 European platform organizations, this study identifies five meta-organizational orchestration mechanisms and develops an empirically grounded model that explains how the focal firm orchestrates value creation with a platform-based circular business model. This study advances the existing knowledge on orchestration mechanisms in platform-based meta-organizations in a circular economy context and highlights novel implications for theory and practice.
Conference Paper
Full-text available
Digital platforms in the gig economy (e.g., Uber, Amazon Mechanical Turk) increasingly use algorithm-based mechanisms to manage workers. One mechanism is algorithmic replacing through which under-performing workers are automatically excluded from platform activities and substituted with other workers. Although algorithmic replacing is an effective means to assure adequate worker output, it causes algoactivism, which refers to worker resistance to the employed algorithms. Drawing on algo-rithmic management literature, we examined algoactivism using topic modeling and grounded theory on driver posts from Uberpeople.net to identify and classify algoactivistic practices in ridesharing. Our results reveal eleven algoactivistic practices that can be arranged and depicted in a "Categories of Algoactivistic Practices in Ridesharing" (CAR) model. Overall, our study expands the literature on al-gorithmic management by shedding light on algoactivism as a crucial worker reaction to algorithmic management and provides practitioners with valuable insights to deal with the phenomenon to ascertain a fair working environment.
Article
The use of algorithms systems that manage employees is now common in many industries. Trucking is no exception since freight companies need sophisticated systems to comply with legal obligations regarding driving time limitations. However, these systems often go far beyond simple GPS tracking, with other advanced surveillance devices being linked to performance management of truck drivers. Despite the growing literature emphasizing how the opacity of these systems may lead to perceived unjust automated decisions, no studies has empirically examined the relationships between algorithmic management transparency on truck driver attitudes. Hence, this study aims to quantitatively investigate the effect of the transparency of two algorithmic management functions (i.e., surveillance and performance management) on distributive and procedural justice. Moreover, considering that this industry faces an extraordinarily high turnover rate, the indirect relationship between algorithmic management transparency and driver's intention to quit through the mediating role of perceived justice will also be examined. Data were collected from 110 respondents via online communities of truck drivers. The results show that the transparency of algorithmic surveillance is positively related to procedural justice whereas the transparency of algorithmic performance management is positively related to distributive justice. Furthermore, our results show that procedural justice mediates the negative relationship between the transparency of algorithmic surveillance and intention to quit and that distributive justice mediates the negative relationship between the transparency of algorithmic performance management and intention to quit. Implications for theory and practice are discussed.
Article
This panel aims to highlight the importance of digital platforms to information science (IS) and in doing this to help connect IS scholars to the important work from other intellectual communities when conducting studies on platforms. The pervasiveness of digital platforms in our lives has encouraged scholars across multiple disciplines to turn their attention to these socio‐technical arrangements. And, over the past two‐plus decades, scholars have advanced multiple conceptualizations and some initial theorizing about digital platforms. This noted, platform‐oriented literature in IS is less focused, and more importantly, does not connect to the conversations on this topic developed in other intellectual communities. To begin connecting IS scholars with platform scholarship outside of our field, this panel brings together IS scholars who have focused their scholarly attention to studying platforms in order to share their research and discuss how their studies could be inspired by and contribute to the platform literature.
Article
Full-text available
Algorithmic control is expanding in various domains with the advances in programming algorithms, the continuous increase in hardware computing power, larger amounts of available fine-grained data, and an increasing number of organizations exercising remote work. Scholars and practitioners in human resource management posit that organizations’ adoption of algorithms as a substitute for or supplement to traditional rational control mechanisms to direct, discipline, and evaluate workers might increase the objectivity and transparency of worker-related decision-making processes and, therefore, reduce the power asymmetry and inequality within organizations. This discussion commentary argues that the underlying assumptions of the higher objectivity and transparency of algorithms in organizational control are very strong, and current evidence does not support them. There is also evidence of large variation in organizations’ adoption of algorithmic control due to their current technical, structural, and human capital resources, which further blurs the predicted outcomes. Evidence also exists for an over-reliance on algorithmic suggestions by managers to circumvent accountability. Adopting algorithmic control must therefore be conducted with serious precautions. This article proposes that overestimation of objectivity and transparency, and large variation in organizations’ adoption of AC (including the lack of technical and managerial knowledge of the underlying mechanisms of learning algorithms in some organizations, and the complete abandonment of human intuitive judgment and reasoning in others) could worsen the power asymmetry and inequality within organizations by increasing the opacity of decisions, systematic biases, discriminatory classification, and violation of worker privacy.
Article
This paper argues that repeated concerns about the nature of academic accounting research should be taken seriously. It proposes engaged research as a means of deepening our social license. Anchored in the argument of Chua and Fiedler (2022), engaged research is characterised as follows. It engages with complex relations of practice, is a crosscultural journey and it co‐produces research on matters that matter (to practitioners, regulators, academics, communities). The paper outlines how accounting research could engage with a contemporary business landscape characterised by pervasive digitisation as well as new risks and opportunities brought about by climate change.
Article
Though gig platforms increasingly employ algorithmic management and monitoring, the benefits and costs of platform monitoring and how these correspond to gig workers’ cognitive work engagement are not well understood. Drawing on supervisor monitoring literature and conservation of resource theory, this study proposes two types of gig platform monitoring (observational monitoring and interactional monitoring) and builds a moderated mediation model to analyze how these platform monitoring methods impact gig workers’ cognitive work engagement. PLS-SEM analysis of data from 269 time-lagged surveys from gig workers shows that, through the mediation of affective trust, observational monitoring is negatively and interactional monitoring is positively related to cognitive work engagement. Furthermore, interactional but not observational monitoring is positively related to cognitive work engagement through the mediation of affective commitment, and method control can enhance the positive relationship between affective trust and cognitive work engagement. These findings contribute to the literature on the nature of gig platform monitoring and its impacts on gig workers’ cognitive work engagement by differentiating observational and interactional monitoring. Gig platform managers will benefit from knowing that interactional but not observational monitoring can promote gig workers’ work engagement.
Article
Full-text available
In the platform economy, digital labor platforms (DLPs) enable remote work by intermediating transactions between platform-dependent entrepreneurs (PDEs) and clients. Information asymmetry arises, however, because PDEs engaged in remote work often are more informed about their abilities than clients are, which may lead to the “lemons problem” and cause market failure. Applying transaction cost economics, we develop a framework that incorporates propositions about what conditions allow different forms of trust to emerge among PDEs, DLPs, and clients. We also delineate how building and sustaining trust should help mitigate market failure threats on DLPs, and, in turn, improve outcomes for PDEs, platforms, and clients. Our framework makes important conceptual extensions by highlighting the relevance of information asymmetry, linking trust development within the DLP-facilitated service triad, and underscoring transaction frequency’s importance, which has been largely overlooked in extant literature. It also provides practical implications for remote work involving PDEs, DLPs, and clients.
Article
We present a model of phenomenon-driven theorizing that may guide researchers in bridging art and science as they engage in pure theory development. The model seeks to leverage activities related to discovery and prescience (conceiving the theory), imagination and logic (constructing the theory), as well as storytelling and scripting (communicating the theory). We illustrate the model through a reflective inquiry into our recent work on data network effects. The model represents a balanced take on theory development and is useful for tempering the natural tendency in academia to drift toward the science of theorizing at the expense of the art of theorizing. © 2021, Association for Information Systems. All rights reserved.
Conference Paper
Full-text available
The gig-economy literature is rife with conflicting accounts of autonomy and empowerment versus exploitation and marginalization. To understand such contradictions, it is necessary to measure perceptions of algorithmic autonomy-support (PAAS); yet no validated instruments exist. To address this gap, we develop a theoretically-based measure for PAAS using Mackenzie et al.'s (2011) well-cited scale development process. To execute our scale development process, interviews were conducted with Uber drivers to support item generation; this was followed by content-validation with subject matter experts to develop and validate our instrument. Lastly, statistical validation was conducted using data collected from a total sample of 435 Uber drivers. The results of our survey confirm that: (i) PAAS is a second-order formative measure with four first-order reflective constructs; (ii) our 13-item scale demonstrates adequate psychometric properties; and (iii) PAAS is positively, and significantly, related to perceived organizational support and job satisfaction. Research contributions and applications are discussed.
Conference Paper
Full-text available
The gig-economy literature is rife with conflicting accounts of autonomy and empowerment versus exploitation and marginalization. To understand such contradictions, it is necessary to measure perceptions of algorithmic autonomy-support (PAAS); yet no validated instruments exist. To address this gap, we develop a theoretically-based measure for PAAS using Mackenzie et al.'s (2011) well-cited scale development process. To execute our scale development process, interviews were conducted with Uber drivers to support item generation; this was followed by content-validation with subject matter experts to develop and validate our instrument. Lastly, statistical validation was conducted using data collected from a total sample of 435 Uber drivers. The results of our survey confirm that: (i) PAAS is a second-order formative measure with four first-order reflective constructs; (ii) our 13-item scale demonstrates adequate psychometric properties; and (iii) PAAS is positively, and significantly, related to perceived organizational support and job satisfaction. Research contributions and applications are discussed.
Article
Full-text available
Artificial Intelligence and algorithmic technologies support or even automate a large variety of human resource management (HRM) activities. This affects a range of stakeholders with different, partially conflicting perspectives on the opacity and transparency of algorithm-based HRM. In this paper, we explain why opacity is a key characteristic of algorithm-based HRM, describe reasons for opaque algorithm-based HRM, and highlight the implications of opacity from the perspective of the main stakeholders involved (users, affected people, deployers, developers, and regulators). We also review strategies to reduce opacity and promote transparency of algorithm-based HRM (technical solutions, education and training, regulation and guidelines), and emphasize that opacity and transparency in algorithm-based HRM can simultaneously have beneficial and detrimental consequences that warrant taking a multi-stakeholder view when considering these consequences. We conclude with a research agenda highlighting stakeholders' interests regarding opacity, strategies to reduce opacity, and consequences of opacity and transparency in algorithm-based HRM. Keywords: explainable Artificial Intelligence, Explainability, Opacity, Transparency, Human Resource Management
Article
Full-text available
Over the past three decades, platform competition—the competition between firms that facilitate transactions and govern interactions between two or more distinct user groups who are connected via an indirect network—has attracted significant interest from the fields of management and organizations, information systems, economics, and marketing. Despite common interests in research questions, methodologies, and empirical contexts by scholars from across these fields, the literature has developed mostly in isolated fashion. This article offers a systematic and interdisciplinary review of the literature on platform competition by analyzing a sample of 333 articles published between 1985 and 2019. The review contributes by (a) documenting how the literature on platform competition has evolved; (b) outlining four themes of shared scholarly interest, including how network effects generate “winner-takes-all” dynamics that influence strategies, such as pricing and quality; how network externalities and platform strategy interact with corporate-level decisions, such as vertical integration or diversification into complementary goods; how heterogeneity in the platform and its users influences platform dynamics; and how the platform “hub” orchestrates value creation and capture in the overall ecosystem; and (c) highlighting several areas for future research. The review aims to facilitate a broader understanding of the platform competition research that helps to advance our knowledge of how platforms compete to create and capture value.
Article
Full-text available
Some of the world's most profitable firms own platforms that exhibit network effects. A platform exhibits network effects if, the more that people use it, the more valuable it becomes to each user. Theorizing about the value perceived by users of a platform that exhibits network effects has traditionally focused on direct and indirect network effects. In this paper, we theorize about a new category of network effects-data network effects- that has emerged from advances in artificial intelligence and the growing availability of data. A platform exhibits data network effects if, the more that the platform learns from the data it collects on users, the more valuable the platformbecomes to each user. We argue that there is a positive direct relationship between the artificial intelligence capability of a platform and the value perceived in the platform by its users-a relationship that is moderated by platform legitimation, data stewardship, and user-centric design.
Article
Full-text available
Drawing on interviews with 77 high-performing eBay business sellers in France and Belgium, this article investigates the power asymmetries generated by customers’ evaluations in online work settings. Sellers revealed a high degree of sensitivity to negative reviews, which, while infrequent, triggered feelings of anxiety and vulnerability. Their accounts exposed power asymmetries at two levels: the transactional level between sellers and customers and the governance level between sellers and eBay. Our findings highlight three main mechanisms underlying power asymmetries in this context. First, online customer evaluations have created a new form of employee monitoring in which power is exercised through the construction of visibility gaps between buyers and sellers and through an implicit coalition between buyers and the platform owner, who join together in the evaluation procedures. Second, by mediating and objectifying relations, algorithms reproduce power asymmetries among the different categories of actors, thereby constraining human agency. Third, online customer evaluations prompt sellers to exploit their practical knowledge of the algorithm to increase their agency. Through the lived experience of working for an algorithm, our findings contribute new understandings of power and agency in online work settings.
Article
Full-text available
This article evaluates the job quality of work in the remote gig economy. Such work consists of the remote provision of a wide variety of digital services mediated by online labour platforms. Focusing on workers in Southeast Asia and Sub-Saharan Africa, the article draws on semi-structured interviews in six countries (N = 107) and a cross-regional survey (N = 679) to detail the manner in which remote gig work is shaped by platform-based algorithmic control. Despite varying country contexts and types of work, we show that algorithmic control is central to the operation of online labour platforms. Algorithmic management techniques tend to offer workers high levels of flexibility, autonomy, task variety and complexity. However, these mechanisms of control can also result in low pay, social isolation, working unsocial and irregular hours, overwork, sleep deprivation and exhaustion.
Article
Full-text available
In the last few years leading-edge research from information systems, strategic management, and economics have separately informed our understanding of platforms and infrastructures in the digital age. Our motivation for undertaking this special issue rests in the conviction that it is significant to discuss platforms and infrastructures concomitantly, while enabling knowledge from diverse disciplines to cross-pollinate to address critical, pressing policy challenges and informstrategic thinking across both social and business spheres. In this editorial,we reviewkey insights from the literature on digital infrastructures and platforms, present emerging research themes, highlight the contributions developed from each of the six articles in this special issue, and conclude with suggestions for further research.
Article
Full-text available
Sharing economy platforms combine organizational and market mechanisms in innovative ways to gain competitive advantages over incumbents. These mechanisms are combined on two key dimensions: tight or loose control over participants, and high or low rivalry between participants. We call the resulting four sharing economy models “Franchiser,” “Principal,” “Chaperone” and “Gardener.” Each model focuses on a different value proposition and strategic intent, but they all exploit the growing fluidity of organizational boundaries. Understanding the sharing economy in these terms enables businesses to identify, and respond to, the threats and opportunities provided by sharing economy platforms. The lessons learned from our research will help businesses to engage with the sharing economy.
Article
Full-text available
As digital platforms are transforming almost every industry today, they are slowly finding their way into the mainstream information systems (ISs) literature. Digital platforms are a challenging research object because of their distributed nature and intertwinement with institutions, markets and technologies. New research challenges arise as a result of the exponentially growing scale of platform innovation, the increasing complexity of platform architectures and the spread of digital platforms to many different industries. This paper develops a research agenda for digital platforms research in IS. We recommend researchers seek to (1) advance conceptual clarity by providing clear definitions that specify the unit of analysis, degree of digitality and the sociotechnical nature of digital platforms; (2) define the proper scoping of digital platform concepts by studying platforms on different architectural levels and in different industry settings; and (3) advance methodological rigour by employing embedded case studies, longitudinal studies, design research, data-driven modelling and visualisation techniques. Considering current developments in the business domain, we suggest six questions for further research: (1) Are platforms here to stay? (2) How should platforms be designed? (3) How do digital platforms transform industries? (4) How can data-driven approaches inform digital platforms research? (5) How should researchers develop theory for digital platforms? and (6) How do digital platforms affect everyday life?
Article
Full-text available
Advancements in Internet and digital technologies have enabled a new work form of open sourcing, which we refer to as the crowdsourcing work environment (CSWE). This new form of work has the potential to disrupt and transform the nature of traditional work. However, our understanding of this new work form is still in its incipient stage. To enhance our understanding, this study captures crowdworkers’ perceptions to explore the characteristics of the crowdworkers, crowdsourcing jobs, and the crowdwork environment that collectively drive the crowdworkers to participate in open source work. Guided by the job characteristic theory and work value perspectives, we used the revealed causal mapping method to analyze narratives by 55 crowdworkers registered on Amazon Mechanical Turk (MTurk). Our data analysis uncovered nine main constructs, 22 key concepts, and 815 causal-effect linkages surrounding CSWE that could guide our theoretical understanding of this emerging phenomenon. Individual needs and the crowdwork context emerged as the major factors motivating individuals’ initial participation in CSWE, but we found crowdsourcing task characteristics (e.g., job autonomy, task variety, task significance, task instruction, and task compensation) and a digitally enabled environment (e.g., system affordance and MTurk governance) to shape crowdworkers’ continued participation in CSWE. The findings suggest several promising research streams, including the psychological factors (i.e., personal growth needs and work values) and social outcomes (i.e., empowerment or exploitation of crowdworkers) for examining the psychology and sociology of crowdsourcing work.
Article
Full-text available
Big data and sophisticated algorithms enable software to handle increasingly complex tasks, such as detecting fraud, optimizing logistics routes, and even driving cars. Beyond technical tasks, algorithms enable new ways to organize work. In this article, I suggest a distinction of optimizing-oriented and open-ended systems leveraging big data and examine how they are shaping organizational design. The optimizing-oriented systems, typically based on numerical data, enable smarter control of well-defined tasks, including algorithmic management of human work. Open-ended systems, often based on textual data or visualizations, can provide answers to a broad range of managerial questions relevant to effective organizing, thereby enabling smarter and more responsive definition of tasks and allocation of resources and effort. Algorithms processing conversations that naturally take place in organizations can form ‘computer augmented transparency’, creating a host of potential benefits, but also threats. These developments are leading to a wave of innovation in organizational design and changes to institutionalized norms of the workplace.
Article
Full-text available
Algorithms, once obscure objects of technical art, have lately been subject to considerable popular and scholarly scrutiny. What does it mean to adopt the algorithm as an object of analytic attention? What is in view, and out of view, when we focus on the algorithm? Using Niklaus Wirth's 1975 formulation that “algorithms + data structures = programs” as a launching-off point, this paper examines how an algorithmic lens shapes the way in which we might inquire into contemporary digital culture.
Article
Full-text available
Rapid and pervasive digitization of innovation processes and outcomes has upended extant theories on innovation management by calling into question fundamental assumptions about the definitional boundaries for innovation, agency for innovation, and the relationship between innovation processes and outcomes. There is a critical need for novel theorizing on digital innovation management that does not rely on such assumptions and draws on the rich and rapidly emerging research on digital technologies. We offer suggestions for such theorizing in the form of four new theorizing logics, or elements, that are likely to be valuable in constructing more accurate explanations of innovation processes and outcomes in an increasingly digital world. These logics can open new avenues for researchers to contribute to this important area. Our suggestions in this paper, coupled with the six research notes included in the special issue on digital innovation management, seek to offer a broader foundation for reinventing innovation management research in a digital world.
Article
Full-text available
In some online labor markets, workers are paid by the task, choose what tasks to work on, and have little or no interaction with their (usually anonymous) buyer/employer. These markets look like true spot markets for tasks rather than markets for employment. Despite appearances, we find via a field experiment that workers act more like parties to an employment contract: workers quickly form wage reference points and react negatively to proposed wage cuts by quitting. However, they can be mollified with "reasonable" justifications for why wages are being cut, highlighting the importance of fairness considerations in their decision making. We find some evidence that "unreasonable" justifications for wage cuts reduce subsequent work quality. We also find that not explicitly presenting the worker with a decision about continuing to work eliminates "quits," with no apparent reduction in work quality. One interpretation for this finding is that workers have a strong expectation that they are party to a quasi-employment relationship where terms are not changed, and the default behavior is to continue working.
Article
Full-text available
This article considers the issue of opacity as a problem for socially consequential mechanisms of classification and ranking, such as spam filters, credit card fraud detection, search engines, news trends, market segmentation and advertising, insurance or loan qualification, and credit scoring. These mechanisms of classification all frequently rely on computational algorithms, and in many cases on machine learning algorithms to do this work. In this article, I draw a distinction between three forms of opacity: (1) opacity as intentional corporate or state secrecy, (2) opacity as technical illiteracy, and (3) an opacity that arises from the characteristics of machine learning algorithms and the scale required to apply them usefully. The analysis in this article gets inside the algorithms themselves. I cite existing literatures in computer science, known industry practices (as they are publicly presented), and do some testing and manipulation of code as a form of lightweight code audit. I argue that recognizing the distinct forms of opacity that may be coming into play in a given application is a key to determining which of a variety of technical and non-technical solutions could help to prevent harm.
Article
Full-text available
Online labor markets are web-based platforms that enable buyers to identify and contract for IT services with service providers using Buyer-Determined (BD) auctions. BD auctions in online labor markets either follow an open or a sealed bid format. We compare open and sealed bid auctions in online labor markets to identify which format is superior in terms of obtaining more bids and a higher buyer surplus. Our theoretical analysis suggests that the relative advantage of open versus sealed bid auctions hinges on the role of reducing service providers’ valuation uncertainty (difficulty in assessing the cost to execute a project) and competition uncertainty (difficulty in assessing the intensity of the competition from other service providers), which largely depends on the relative importance of the common value (versus the private value) component of the auctioned IT services, calling for an empirical investigation to compare open and sealed bid auctions. Based on a unique dataset of 71,437 open bid auctions and 7,499 sealed bid auctions posted by 21,799 buyers at a leading online labor market, we find that, on average, albeit sealed bid auctions attract 18.4% more bids, open bid auctions offer buyers $10.87 higher surplus. Furthermore, open bid auctions are 55.3% more likely to result in a buyer’s selection of a certain service provider, 22.1% more likely to reach a contract (conditional on the buyer’s making a selection) with a provider, and they generate higher buyer satisfaction. In contrast to conventional wisdom that “the more bids the better” and industry practice of treating sealed bid auctions as a premium feature, our results suggest that the buyer surplus gained from the reduction in valuation uncertainty enabled by open bid auctions outweighs the buyer surplus gained from the higher competition uncertainty in sealed bid auctions, which renders open bid auctions a superior auction design in online labor markets.
Article
Full-text available
In this essay, I begin by identifying the reasons that automation has not wiped out a majority of jobs over the decades and centuries. Automation does indeed substitute for labor—as it is typically intended to do. However, automation also complements labor, raises output in ways that leads to higher demand for labor, and interacts with adjustments in labor supply. Journalists and even expert commentators tend to overstate the extent of machine substitution for human labor and ignore the strong complementarities between automation and labor that increase productivity, raise earnings, and augment demand for labor. Changes in technology do alter the types of jobs available and what those jobs pay. In the last few decades, one noticeable change has been a "polarization" of the labor market, in which wage gains went disproportionately to those at the top and at the bottom of the income and skill distribution, not to those in the middle; however, I also argue, this polarization is unlikely to continue very far into future. The final section of this paper reflects on how recent and future advances in artificial intelligence and robotics should shape our thinking about the likely trajectory of occupational change and employment growth. I argue that the interplay between machine and human comparative advantage allows computers to substitute for workers in performing routine, codifiable tasks while amplifying the comparative advantage of workers in supplying problem-solving skills, adaptability, and creativity.
Conference Paper
Full-text available
Software algorithms are changing how people work in an ever-growing number of fields, managing distributed human workers at a large scale. In these work settings, human jobs are assigned, optimized, and evaluated through algorithms and tracked data. We explored the impact of this algorithmic, data-driven management on human workers and work practices in the context of Uber and Lyft, new ridesharing services. Our findings from a qualitative study describe how drivers responded when algorithms assigned work, provided informational support, and evaluated their performance, and how drivers used online forums to socially make sense of the algorithm features. Implications and future work are discussed.
Article
Full-text available
Alternatives to the archetypal model of full-time regular employment are now both prevalent and wide-ranging. Over a fifth of U.S. workers, and even more globally, now perform economic work under arrangements that differ from full-time regular employment. Yet most of our management and social science notions about economic work are based on the full-time employment model. We know relatively little about the operation and consequences of alternative arrangements in part because while these arrangements vary considerably, they are commonly grouped together for research purposes using existing classification systems. We outline an inclusive classification system that distinguishes clearly between employment and its alternatives. It also distinguishes among the alternatives themselves by grouping work arrangements into categories that share common properties and that are distinct from each other in ways that matter for practice and for research. The classification system is based on distinctions about the sources and extent of control over the work process, the contractual nature of the work relationship, and the parties involved in the work relationship. Our classification system is both informed by and reflects the legal distinctions among these categories. We explore implications of our system for research and theory development.
Article
Full-text available
Today, digital data are captured through a variety of devices that have the ability to monitor the minutiae of an individual’s everyday life. These data are often processed by algorithms, which support (or drive) decisions (termed ‘algorithmic decision-making’ in this article). While the strategic value of these data (and subsequent analysis) for businesses is unquestionable, the implications for individuals and wider society are less clear. Therefore, in this Viewpoint article we aim to shed light on the tension between businesses – that increasingly profile customers and personalize products and services – and individuals, who, as McAfee and Brynjolfsson (2012, p. 5) suggest, are ‘walking data generators’ but are often unaware of how the data they produce are being used, and by whom and with what consequences. Issues associated with privacy, control and dependence arise, suggesting that social and ethical concerns related to the way business is strategically exploiting digitized technologies that increasingly support our everyday activities should be brought to the fore and thoughtfully discussed. In this article we aim to lay a foundation for this discussion in the IS community and beyond.
Article
Full-text available
Big data and the mechanisms by which it is produced and disseminated introduce important changes in the ways information is generated and made relevant for organizations. Big data often represents miscellaneous records of the whereabouts of large and shifting online crowds. It is frequently agnostic, in the sense of being produced for generic purposes or purposes different from those sought by big data crunching. It is based on varying formats and modes of communication (e.g., texts, image and sound), raising severe problems of semiotic translation and meaning compatibility. Crucially, the usefulness of big data rests on their steady updatability, a condition that reduces the time span within which this data is useful or relevant. Jointly, these attributes challenge established rules of strategy making as these are manifested in the canons of procuring structured information of lasting value that addresses speci␣c and long-term organizational objectives. The developments underlying big data thus seem to carry important implications for strategy making, and the data and information practices with which strategy has been associated. We conclude by placing the understanding of these changes within the wider social and institutional context of longstanding data practices and the significance they carry for management and organizations.
Conference Paper
Full-text available
Humans can perform many tasks with ease that remain difficult or impossible for computers. Crowdsourcing platforms like Amazon's Mechanical Turk make it possible to harness human-based computational power at an unprecedented scale. However, their utility as a general-purpose computational platform remains limited. The lack of complete automation makes it difficult to orchestrate complex or interrelated tasks. Scheduling more human workers to reduce latency costs real money, and jobs must be monitored and rescheduled when workers fail to complete their tasks. Furthermore, it is often difficult to predict the length of time and payment that should be budgeted for a given task. Finally, the results of human-based computations are not necessarily reliable, both because human skills and accuracy vary widely, and because workers have a financial incentive to minimize their effort. This paper introduces AutoMan, the first fully automatic crowdprogramming system. AutoMan integrates human-based computations into a standard programming language as ordinary function calls, which can be intermixed freely with traditional functions. This abstraction lets AutoMan programmers focus on their programming logic. An AutoMan program specifies a confidence level for the overall computation and a budget. The AutoMan runtime system then transparently manages all details necessary for scheduling, pricing, and quality control. AutoMan automatically schedules human tasks for each computation until it achieves the desired confidence level; monitors, reprices, and restarts human tasks as necessary; and maximizes parallelism across human workers while staying under budget.
Article
Full-text available
An integrative framework is proposed to advance management research on technological platforms, bridging two theoretical perspectives: economics, which sees platforms as double-sided markets, and engineering design, which sees platforms as technological architectures. While the economic perspective informs our understanding of platform competition, the engineering design perspective informs our view of platform innovation. The article argues that platforms can be usefully conceptualized as evolving organizations or meta-organizations that: (1) federate and coordinate constitutive agents who can innovate and compete; (2) create value by generating and harnessing economies of scope in supply or/and in demand; and (3) entail a modular technological architecture composed of a core and a periphery. In support of this conceptualization, a classification system is presented, indicating that technological platforms appear in a variety of organizational forms: within firms, across supply chains, and across industry innovation ecosystems. As an illustration, the framework is then applied to derive a simple model highlighting patterns of interaction between platform innovation and competition, yielding hypotheses that could be tested empirically by future scholars.
Article
Full-text available
For all its richness and potential for discovery, qualitative research has been critiqued as too often lacking in scholarly rigor. The authors summarize a systematic approach to new concept development and grounded theory articulation that is designed to bring “qualitative rigor” to the conduct and presentation of inductive research.
Article
Full-text available
The growth of precarious work since the 1970s has emerged as a core contemporary concern within politics, in the media, and among researchers. Uncertain and unpredictable work contrasts with the relative security that characterized the three decades following World War II. Precarious work constitutes a global challenge that has a wide range of consequences cutting across many areas of concern to sociologists. Hence, it is increasingly important to understand the new workplace arrangements that generate precarious work and worker insecurity. A focus on employment relations forms the foundation of theories of the institutions and structures that generate precarious work and the cultural and individual factors that influence people's responses to uncertainty. Sociologists are well-positioned to explain, offer insight, and provide input into public policy about such changes and the state of contemporary employment relations.
Article
The widespread implementation of algorithmic technologies in organizations prompts questions about how algorithms may reshape organizational control. We use Edwards’ (1979) perspective of “contested terrain,” wherein managers implement production technologies to maximize the value of labor and workers resist, to synthesize the interdisciplinary research on algorithms at work. We find that algorithmic control in the workplace operates through six main mechanisms, which we call the “6 Rs”—employers can use algorithms to direct workers by restricting and recommending, evaluate workers by recording and rating, and discipline workers by replacing and rewarding. We also discuss several key insights regarding algorithmic control. First, labor process theory helps to highlight potential problems with the largely positive view of algorithms at work. Second, the technical capabilities of algorithmic systems facilitate a form of rational control that is distinct from the technical and bureaucratic control used by employers for the past century. Third, employers’ use of algorithms is sparking the development of new algorithmic occupations. Finally, workers are individually and collectively resisting algorithmic control through a set of emerging tactics we call algoactivism. These insights sketch the contested terrain of algorithmic control and map critical areas for future research.
Article
Learning algorithms, technologies that generate responses, classifications, or dynamic predictions that resemble those of a knowledge worker, raise important research questions for organizational scholars related to work and organizing. We suggest that such algorithms are distinguished by four consequential aspects: black-boxed performance, comprehensive digitization, anticipatory quantification, and hidden politics. These aspects are likely to alter work and organizing in qualitatively different ways beyond simply signaling an acceleration of long-term technology trends. Our analysis indicates that learning algorithms will transform expertise in organizations, reshape work and occupational boundaries, and offer novel forms of coordination and control. Thus, learning algorithms can be considered performative due to the extent to which their use can shape and alter work and organizational realities. Their rapid deployment requires scholarly attention to societal issues such as the extent to which the algorithm is authorized to make decisions, the need to incorporate morality in the technology, and their digital iron-cage potential.
Article
Learning algorithms, technologies that generate responses, classifications, or dynamic predictions that resemble those of a knowledge worker, raise important research questions for organizational scholars related to work and organizing. We suggest that such algorithms are distinguished by four consequential aspects: black-boxed performance, comprehensive digitization, anticipatory quantification, and hidden politics. These aspects are likely to alter work and organizing in qualitatively different ways beyond simply signaling an acceleration of long-term technology trends. Our analysis indicates that learning algorithms will transform expertise in organizations, reshape work and occupational boundaries, and offer novel forms of coordination and control. Thus, learning algorithms can be considered performative due to the extent to which their use can shape and alter work and organizational realities. Their rapid deployment requires scholarly attention to societal issues such as the extent to which the algorithm is authorized to make decisions, the need to incorporate morality in the technology, and their digital iron-cage potential.
Article
We examine how the entry of gig-economy platforms influences local entrepreneurial activity. On the one hand, such platforms may reduce entrepreneurial activity by offering stable employment for the unemployed and underemployed. On the other hand, such platforms may enable entrepreneurial activity by offering work flexibility that allows the entrepreneur to redeploy resources strategically in order to pursue the nascent venture. To resolve this tension, we examine the entry of the ridesharing platform Uber X into local areas. We use two measures of entrepreneurial activity: crowdfunding campaign launches at Kickstarter, the world’s largest reward-based crowdfunding platform, and levels of self-employment from the Current Population Survey. Results indicate a negative and significant relationship between platform entry and both measures of entrepreneurial activity. Importantly, the effect manifests primarily amongst unsuccessful Kickstarter campaigns and unincorporated entrepreneurial ventures, suggesting that gig-economy platforms predominantly reduce lower quality entrepreneurial activity, seemingly by offering viable employment for the unemployed and underemployed. These relationships are corroborated with a first-hand survey conducted with gig-economy service providers. This paper was accepted by Anandhi Bharadwaj, information systems.