“The only real choice I have is to reject offers or log out, which doesn’t feel like that much of a freedom really”: Migrant Labour, Algorithmic Control and Social Network in the Platform Economy

Technical Summary: Migrant Labour, Algorithmic Control and Social Network in the Platform Economy

Problem Statement
The paper addresses a critical gap in labour scholarship regarding how migrant workers, particularly those with temporary legal status, experience and navigate "algorithmic management" within the platform economy. While existing literature acknowledges that migrants are disproportionately represented in gig work (e.g., food delivery, ride-sharing) and that algorithmic opacity creates "algorithmic precarity," there is limited understanding of how this specific cohort interprets, anticipates, and resists these opaque digital control systems. The study investigates whether migrant workers' reliance on social and ethnic networks serves as a mechanism to mitigate the isolation and exploitation inherent in platform work, or if these networks merely reproduce existing hierarchies and exclusions.

Methodology
The research employs a qualitative case study approach focused on migrant food delivery riders in Melbourne, Australia.

• Data Collection: Semi-structured interviews were conducted with 55 migrant gig workers (17 ridesharing drivers and 38 food delivery riders) across platforms including Uber, Uber Eats, DoorDash, Hungry Panda, and Easi. Additionally, six key stakeholders (government officials, union leaders, and industry insiders) were interviewed. For this specific paper, the analysis is restricted to the 38 food delivery riders.
• Sampling: Participants were recruited via purposive sampling from the Melbourne Gig Workers' Hub and a local bike repair shop. The cohort is diverse, hailing from China, Taiwan, India, Pakistan, Colombia, Mexico, Malaysia, and Indonesia. Most participants hold student visas, with a high proportion holding undergraduate degrees or higher.
• Analysis: Data was analyzed using a thematic approach (Braun & Clarke, 2006) combined with an abductive approach, allowing for a navigation between existing theory and participant accounts. Transcripts were coded to identify patterns regarding algorithmic understanding, social network usage, and resistance tactics.

Key Findings
The study reveals a complex interplay between algorithmic control, migrant status, and ethnic social networks:

1. "Illusory" Flexibility and Constrained Agency:
   Migrant workers possess a high degree of "procedural literacy" regarding platform algorithms, understanding how ratings, location data, and acceptance rates influence order allocation. However, this knowledge does not translate into meaningful autonomy. Workers describe their agency as confined to a binary choice: accepting an algorithmically assigned task or "logging out." This creates a condition of "illusory flexibility," where the appearance of independence masks a reality of coercive control where the only true exit strategy is disengagement.

2. Algorithmic Opacity and Hybrid Control:
   Platforms utilize "algorithmic opacity" to maintain power. Workers can infer patterns (e.g., proximity to restaurants increases order volume, high rejection rates trigger system adjustments) but cannot fully predict or challenge decision-making. The study identifies a "hybrid management" model, particularly on platforms like Hungry Panda, where algorithmic tracking is supplemented by human managers who physically verify rider compliance (uniforms, routes), creating a surveillance regime that intensifies self-discipline without explicit notification of penalties.

3. Ethnic Networks as Counter-Infrastructures:
   Migrant workers utilize WhatsApp and WeChat groups as informal "counter-infrastructures" to combat algorithmic opacity. These networks function as collective intelligence systems where workers share real-time data on "hotspots," order volumes, platform updates, and strategies for managing customer ratings or cancellations. This peer-to-peer information sharing allows workers to partially reverse information asymmetries and optimize earnings.

4. The Politics of Belonging and Exclusion:
   Access to these vital networks is uneven and governed by a "politics of belonging." Entry into specific ethnic groups (e.g., Indonesian or Latin American WhatsApp groups) often requires sponsorship or verification by existing members. This creates internal hierarchies where new or unconnected migrants are excluded from critical resources. Furthermore, the study finds that these networks can foster racialised competition rather than solidarity; for instance, Chinese riders expressed anxiety that Indian workers were "taking jobs" or being used by platforms to undermine collective bargaining efforts.

5. Collective Bargaining and Resistance:
   Despite internal divisions, ethnic networks have demonstrated the capacity for collective action. The paper cites a specific instance where Chinese riders on Hungry Panda used their WeChat network to organize a protest against a wage reduction. While the riders negotiated with platform management, the original goal of restoring the unit price to five dollars was not met, as management claimed the price was beyond their control. However, the platform did implement some functional improvements to the app based on the riders' suggestions. The paper notes that such resistance is often limited by ethnic fragmentation, which can hinder broader, cross-ethnic solidarity.

Key Contributions
The paper makes four specific contributions to migration and platform studies:

1. Reconceptualizing Flexibility: It demonstrates that migrant platform workers' procedural literacy does not yield autonomy but rather a constrained agency limited to "accepting or rejecting" tasks, reinforcing the concept of "illusory flexibility."
2. Social Networks as Infrastructure: It identifies WhatsApp and WeChat groups as essential informal counter-infrastructures that enable collective sense-making and limited resistance against algorithmic governance.
3. Uneven Access to Knowledge: It argues that algorithmic knowledge is a scarce resource distributed unevenly within migrant communities, shaped by ethnicity and existing social ties, thereby functioning as both a mechanism of solidarity and exclusion.
4. Racialised Competition: It reveals how algorithmic systems interact with migration status to produce racialised competition among workers, complicating the potential for unified labour resistance.

Significance and Claims
The paper claims that algorithmic governance is not merely a technical phenomenon but is deeply mediated through the social organization of migrant communities. It argues that while ethnic networks provide a vital survival mechanism and a basis for negotiation, they are ambivalent: they offer protection against isolation and opacity but simultaneously reproduce exclusion and racialised competition. The study suggests that understanding the "migration–algorithm nexus" is critical for analyzing the future of platform labour, as the capacity to resist algorithmic control is increasingly dependent on access to these bounded social networks rather than formal labour rights. The author maintains a modest scope, noting that their findings are based on a highly educated cohort of student visa holders in Australia, and calls for future research to explore cross-ethnic alliances and the intersection of gender and migration in platform work.