This paper asks two questions: what do we mean when using the adjective “algorithmic” to describe the quality or nature of control or management at work? And what do we miss about the daily operations of platform power when focusing on algorithms as the culprit behind exacerbated subordination and precarity in the workplace? By answering these questions, the paper contributes to debates on the impact of algorithms in platform-mediated workplaces and, by extension, on the location and nature of platform power.

Drawing on two years of ethnographic fieldwork in the app-based food delivery markets of Berlin, Amsterdam and New York, as well as a comparative analysis of the institutional and regulatory settings in which these markets are embedded, I will advance two empirically substantiated arguments. First, I will argue that we need to better situate our analyses of algorithms at work by attending to what I call the “para-algorithmic”, or that which simultaneous exceeds and frames algorithmic operations. Second, I will contend that, in the context of the gig economy, the most significant resources of power actually reside beyond the platform itself. This invites a critical reassessment of the notion of platform power, one that decenters the algorithm and indeed the digital platform itself.

How has the platformization of neighborhood restaurants and electrification of bicycles reconstituted food delivery work? What work problems have emerged and how have they been addressed by delivery riders within their local and digital storytelling networks? To address these questions, we use digital ethnography of public social media accounts of the two, largest delivery worker communities in New York City combined with in-person observations of these communities and neighborhood fieldwork with food delivery workers, restaurant owners, ebike mechanics, and other stakeholders in the NYC delivery world. We find that the operational costs that platform companies pass onto individual workers are shouldered collectively by delivery workers and their networks. We also find that in the absence of an urban infrastructure that can sustain this industry, crime victimization and occupational injuries have proliferated among this already vulnerable migrant worker population. To meet these conditions of 'liminal precarity' (van Doorn, 2023), delivery workers have developed new forms of DIY urbanism (Douglas, 2018) and vigilance, and a kind of civil society has emerged around migrant delivery work. This study holds theoretical applications for urban communication scholars interested in how local communication ecologies impact politics/civil society, infrastructure, and other aspects of “communicative cities” (Gumpert & Drucker, 2008), as well as practical implications for understanding post-pandemic organizing and mutual aid practices.

The existing literature on algorithmic management mainly focuses on its external consequences for workers. Drawing upon the case of Meituan Waimai, China’s dominant food-delivery platform, this article examines the internal mechanisms that constitute the design of algorithmic management. Grounding my analysis in actual systems and narratives from the production side, I argue that human choices move the development of labor control despite the seeming automaticity of algorithmic management. Three tenets consistently guide the design of algorithmic management at Meituan: the optimal mindset, labor as mechanical resources, and labor as repair. Characterized by a mix of central planning and improvisation, algorithmic management is designed to standardize the labor process without sacrificing the space for flexible production. By highlighting human choices behind the technical surface, the article provides theoretical and methodological lessons for scholars researching algorithmic management and algorithmic power in general.

How can the conditions of computational labor be improved? Studies of digital labor platforms demonstrate how regulations and platform design shape workers’ outcomes. However, prior research tends to overlook how organizational design can influence workers’ experiences. This article draws on 19 months of participant-observation research inside a tech startup that employed 200 computational workers as long-term “team members” integrated into the company’s organizational structure. I first examine how this arrangement ameliorated some of the problems commonly associated with computational labor while leaving others unaddressed. I then turn to the team’s organizational culture of familial love, which strengthened ties between workers and the firm while simultaneously obscuring how the company’s labor practices perpetuated vast inequalities. Although this study suggests how computational labor can be structured in ways that advance organizational goals while simultaneously supporting the dignity of workers, it also reveals the durability of the disparities that characterize tech companies.

The existing literature on the data annotation labor behind artificial intelligence (AI) tends to presume the digital platforms to be the organizational form of the data labor for AI, albeit a geographically dispersed one. In this paper, I revisit this presumption and intend to destabilize the platform-centric understanding of the organization of data work. Taking data annotation labor in China as a point of departure, the paper tackles these questions: 1) What are the organizational forms involved in the labor process of data annotation? 2) What contributes and shapes these organizing practices? The study expands on the organization studies approach and is based on an analysis of the industry reports, ethnographic observations, and interviews. Examining a multiplicity of organizational forms and practices that enable the circulation of data labor, I will make two points. First, the assortment and graft of organizational forms and practices decenter the annotation platforms as the dominant organizing force in the labor process of data annotation. Second, friction and incongruence between different organizational contexts intensify the demand of worker’s communicative labor to ensure the accuracy and quality of the annotation work. Consequently, a series of communication technology, management strategy, and social relations is mobilized to meet the respective communicative and transactional needs, which shapes the organizational practices. There is hardly a platform logic that controls the labor process. This understanding rescues the analytic fixation on the digital platforms and their socio-technical configuration and propels the researchers to examine the organizing logics beyond the platform.