Who’s Doing the Work? The Hidden Labor Behind Delivery Robots

by

Author: Denis Avetisyan

A new analysis reveals that the automation of delivery services isn’t about replacing human workers, but rather shifting and obscuring the labor required to keep these robots running.

The deployment of delivery robots across Seoul’s diverse urban landscapes - from the bustling intersections of central business districts to the carefully planned sidewalks of smart city pilot programs - demonstrates a systemic integration into the existing fabric of pedestrian life, necessitating navigation not merely of space, but of established patterns of human activity and infrastructural challenges like drainage systems and café seating.
The deployment of delivery robots across Seoul’s diverse urban landscapes – from the bustling intersections of central business districts to the carefully planned sidewalks of smart city pilot programs – demonstrates a systemic integration into the existing fabric of pedestrian life, necessitating navigation not merely of space, but of established patterns of human activity and infrastructural challenges like drainage systems and café seating.

This paper argues that delivery robot ‘labor’ is a collectively sustained sociotechnical achievement demanding critical attention to the politics of automation and labor redistribution in urban spaces.

The increasing prevalence of sidewalk delivery robots prompts a reconsideration of fundamental assumptions about labor and automation. This paper, ‘Is Robot Labor Labor? Delivery Robots and the Politics of Work in Public Space’, investigates the often-hidden human effort sustaining these seemingly autonomous systems through ethnographic fieldwork in Seoul’s smart-city districts. We demonstrate that successful robotic delivery is not simply a technological feat, but a distributed sociotechnical achievement reliant on ongoing human coordination and accommodation, effectively reconfiguring-rather than replacing-existing labor practices. How can human-robot interaction research more fully account for the spatial and political dimensions of automation and the resulting shifts in the organization of work?

The Allure of Automated Progress

South Korea is actively positioning robotic delivery systems as more than just logistical tools; they represent a powerful national narrative of technological progress. This ambition extends beyond simple efficiency gains, functioning as a visible symbol of the nation’s commitment to innovation and future-facing industries. The deployment of these robots isn’t solely about streamlining deliveries, but rather a deliberate effort to cultivate a public image of South Korea as a leader in automation and a hub for cutting-edge technology. This strategic framing, supported by substantial investment and proactive policy, aims to inspire both domestic confidence and international recognition, solidifying the nation’s place in the rapidly evolving landscape of global technological advancement.

South Korea’s push towards widespread robotic integration isn’t simply market-driven; it’s powerfully propelled by dedicated, state-led technological development initiatives. These programs strategically fund research, development, and deployment of robotic systems across various sectors, fostering a national ecosystem geared towards automation. Crucially, innovation isn’t hampered by rigid regulations; instead, the nation utilizes innovative ‘Regulatory Sandboxes’ – controlled environments where new technologies can be tested and refined outside of typical legal constraints. This allows for rapid iteration and problem-solving, enabling the safe and efficient integration of robotics into public spaces and daily life, while simultaneously providing real-world data to inform future policy and standards. The combination of proactive funding and flexible regulation demonstrates a deliberate national strategy to establish leadership in the burgeoning field of robotics.

The envisioned robotic futures, while appearing to automate tasks and displace human workers, are in fact deeply embedded within existing urban systems and reliant on a largely unseen human workforce. Research indicates that the implementation of robotic delivery and service systems doesn’t eliminate jobs so much as redistribute them, creating new roles in maintenance, logistics, and system oversight. These robotic networks necessitate robust infrastructure – meticulously mapped routes, dedicated charging stations, and constant data connectivity – all of which require significant human input for construction, upkeep, and monitoring. Consequently, the automation narrative often obscures the persistent and evolving role of human labor, highlighting a shift in employment patterns rather than a complete replacement of workers by machines.

The Hidden Architecture of Automation

Direct observation of delivery robots in operational settings demonstrates substantial human labor is required for continued functionality. This labor extends beyond basic maintenance and includes pre-operational checks, real-time remote assistance for navigation and obstacle avoidance, post-delivery retrieval and recharging, and ongoing data analysis to improve route optimization and identify potential mechanical issues. Observed tasks included manual adjustments to sensor calibration, physical intervention to free robots stuck in challenging terrain, and frequent battery swaps. The consistent need for these interventions indicates that current delivery robot systems are not operating autonomously, but rather as a distributed network reliant on a dedicated human workforce performing tasks often unseen by the public.

The operational demands of delivery robots extend beyond routine maintenance, requiring a level of skilled intervention best described by the Korean concept of Giteukhada. This term denotes admirable competence – a dedication to perfecting a skill through diligent practice and nuanced understanding. Observations reveal that maintaining robot functionality necessitates proactive problem-solving, precise adjustments based on environmental factors, and a deep understanding of the robot’s complex systems. This isn’t simply fixing broken parts; it’s a continuous process of optimization and refinement performed by highly trained technicians, demonstrating that robotic operation relies on a sophisticated skillset and dedicated expertise rather than solely automated processes.

Delivery robot operation is fundamentally a sociotechnical system, meaning its functionality relies not on replacing human labor, but on redistributing it. Observations demonstrate that while robots perform the physical task of delivery, substantial human expertise is required for ongoing operation – including remote monitoring, exception handling for navigation issues, battery swapping, preventative maintenance, and addressing unexpected obstacles. This dependence on human intervention actively challenges the widely promoted narrative of full automation and autonomous systems. Consequently, the implementation of robotic delivery does not eliminate work; rather, it shifts the nature of required labor from direct delivery tasks to specialized roles focused on robot upkeep and management, creating a new division of labor within the logistics network.

The Shifting Sands of Labor

The implementation of delivery robotics is not resulting in net job loss, but rather a redistribution of labor demands. While some roles associated with direct delivery are impacted, new positions are emerging in areas such as robot maintenance, remote monitoring, fleet management, and logistical support. This shift also entails a geographic relocation of work, moving tasks from traditional driving routes to centralized control centers and repair facilities. Data indicates that the overall volume of work remains consistent, but the skill sets and locations required are changing, necessitating workforce adaptation and retraining initiatives to accommodate these evolving demands.

Staged Autonomy refers to the common presentation of robotic systems as fully independent, despite their continuous reliance on human support for tasks such as remote intervention, exception handling, and data labeling. This misrepresentation obscures the ongoing need for a human workforce to maintain and operate these systems, effectively shifting labor from direct service roles to roles focused on monitoring, maintenance, and algorithmic training. While robotic systems may perform the visible work, a significant, often hidden, human component remains essential for their continued function and adaptation to unforeseen circumstances, creating a complex interplay between automated and human labor.

The increasing deployment of delivery robots and similar automated systems is accompanied by a rise in algorithmic oversight of the human labor that supports and manages them. This oversight extends beyond simple task assignment to include performance monitoring and evaluation, often utilizing data collected from the robotic systems themselves and from worker interactions with those systems. Our research indicates this isn’t simply a case of automation replacing workers; rather, existing roles are being reshaped, with algorithmic systems increasingly determining workflow, performance metrics, and, consequently, worker agency. This redistribution of labor is coupled with a shift in control dynamics, as algorithms mediate the relationship between employers and the workforce involved in maintaining and operating these automated solutions.

The Privilege of Automated Presence

The pervasive tendency to accommodate robots in public spaces, termed ‘robot privilege,’ highlights a fundamental asymmetry in how humans navigate shared urban environments. Observations reveal individuals routinely cede right-of-way, offer assistance, and even preemptively adjust their behavior around robots, actions rarely mirrored in interactions between humans themselves. This isn’t simply politeness; the ease with which robots are integrated into pedestrian flows suggests a pre-emptive granting of access and a subtle expectation of robotic mobility. Such accommodation isn’t inherent to the technology, but rather a learned social response, effectively granting robots a default claim to public space and subtly reshaping the norms of urban interaction. The implications extend beyond mere convenience, pointing to a broader renegotiation of social rules and spatial boundaries within increasingly automated environments.

Robot mobility within cities isn’t solely a matter of advanced engineering; it’s deeply entwined with pre-existing urban designs. Current infrastructure – sidewalks, crosswalks, even building accessibility features – inadvertently grants robots a form of privilege, allowing them to navigate spaces designed for humans with relative ease. This isn’t a neutral accommodation; repeated observation suggests it fosters a social expectation that robots have a right-of-way, subtly shifting how people perceive and interact with these machines in shared environments. The effect is a self-reinforcing cycle where infrastructural support legitimizes robotic presence, encouraging further adaptation of public spaces to prioritize robotic movement and ultimately normalizing a hierarchy of access within the urban landscape.

Contemporary “smart city” initiatives often employ dedicated testbeds – controlled urban environments – to showcase robotic technologies, but these spaces do more than simply demonstrate functionality. These carefully constructed settings actively cultivate a perception of robots as natural and seamless components of city life, effectively masking the intricate engineering, data dependencies, and potential disruptions inherent in their operation. Recent research confirms this isn’t merely a process of technological integration; instead, it represents a deliberate social restructuring, subtly reshaping expectations around urban navigation, right-of-way, and the very definition of a public space to accommodate automated agents. By presenting robotic integration as frictionless and inevitable within these testbeds, a specific socio-technical future is actively being rehearsed and normalized, potentially prefiguring broader shifts in how humans interact with – and cede space to – autonomous systems in the wider urban landscape.

The study illuminates how automated systems, such as delivery robots, are not isolated technical achievements but embedded within complex webs of human labor. This echoes Bertrand Russell’s observation: “The difficulty of philosophy is that it attempts to deal with the ultimate questions, not as they appear to the unrefined mind, but as they appear after prolonged reflection.” The seeming autonomy of these robots obscures the ‘prolonged reflection’ – the ongoing human effort in design, maintenance, and logistical support – that underpins their operation. Just as technical debt resembles erosion, the ‘labor’ performed by these robots is not solely their own, but a collective process of ongoing adjustment and concealed human contribution, fundamentally reshaping the politics of work in public spaces.

What Remains to Be Seen

The notion of ‘robot labor’ proves, upon inspection, a misnomer. This work demonstrates that what appears as automation is, in fact, a re-distribution of effort – a shifting of tasks, not an elimination. The system does not erase labor; it merely obscures its provenance. Uptime is not inherent to the device, but a consequence of continuous, often-invisible, human maintenance. The focus on autonomous functionality, then, feels increasingly like a distraction. The question is not whether the robot can work, but who bears the cost of its operation, and what imbalances are thereby created.

Future inquiry must move beyond the mechanics of human-robot interaction and address the politics of it. Stability is an illusion cached by time, a temporary equilibrium maintained by unseen hands. The field requires a reckoning with ‘robot privilege’ – the unearned advantages conferred by automated systems. Consider the latency inherent in every request; latency is the tax every request must pay. To ignore the human infrastructure supporting these devices is to accept a fundamentally incomplete understanding of urban automation.

The trajectory suggests a deepening entanglement of human and machine, not a clean separation. The challenge lies not in perfecting the illusion of autonomy, but in designing sociotechnical systems that acknowledge, and equitably distribute, the enduring reality of human effort. The task isn’t to build robots that replace labor, but to understand how they reshape it-and at what cost.

Original article: https://arxiv.org/pdf/2602.20180.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-02-25 11:46