In June 2026, governments, employers, and worker representatives will come together at the International Labour Organization (ILO) to negotiate what could become the first global standard on decent work in the platform economy. For many, this is a long-awaited recognition that gig and platform work has become central to today’s labour markets. From ride-hailing and delivery to domestic services and data work, platform labour now underpins vast segments of the global economy, as platformization continues to expand across sectors.
However, this moment of recognition is happening within the institutional boundaries of the ILO’s standard-setting process. Tripartite negotiation between states, employers, and worker representatives is bound to produce a compromise. The current draft of the standards (including a binding Convention and a non-binding Recommendation), contained in the Blue Report, reflects this balancing act: extending labor rights while preserving flexibility, acknowledging platform work while avoiding obligations that might disrupt business models, and recognizing new risks while stopping short of regulating how these risks are actually produced.
The draft largely approaches platform work through the lens of existing labor law, relying on familiar categories such as employment status, national jurisdiction, and the notion of workplace as a bounded site.
The standards are also being shaped by a deeper tension. The draft largely approaches platform work through the lens of existing labor law, relying on familiar categories such as employment status, national jurisdiction, and the notion of workplace as a bounded site. But platformization of work is reorganizing these very categories. Work is increasingly controlled through data-driven algorithms, performed across dispersed and often invisible workplaces, and structured through chains of intermediaries (including subcontractors) that fragment responsibility while consolidating control. For many workers, especially in the Global South, these arrangements are further shaped by cross-border outsourcing and migration regimes that complicate access to rights and remedies.
From February to April 2026, the Global Platform Workers Solidarity Project (GPWSP) hosted discussions and strategy sessions with platform workers from across all five continents. These discussions happened in parallel across two broad sectors of platform work: 13 data and other online platform workers from eight different grassroots workers organizations, and 15 platform care workers, including domestic, beauty, nannying, and healthcare workers from nine grassroots workers’ organizations. This article draws upon these conversations and examines workers’ experiences and what they demand from the ILO process and beyond. Their accounts reveal the misalignment between the traditional conceptual frameworks used to regulate work and the forms through which work is now organized.
Redistribution of costs without assignment of responsibility
Across the discussions, one issue surfaced repeatedly: workers are not only underpaid, but are expected to absorb the costs necessary to perform the work itself. This cost allocation reflects a feature of platform business models, which allows platforms to grow without assuming the costs typically associated with expanding a workforce.
In data work, workers are often expected to cover the costs of their own laptops, internet connections, electricity, and often a dedicated workspace. A worker representative from Kenya described the conditions: “Workers are putting in eighteen hours a day and receiving eighteen dollars at the end of the week”, with no compensation for the costs incurred in producing that labor.
In platform care work, the cost burden is differently structured but similarly cumbersome. Workers report paying for training, background checks, uniforms, and transport between jobs. They are also required to comply with client demands that extend beyond agreed tasks, often without compensation, since refusal risks negative ratings or loss of future work. One worker described a recurring situation in which workers are penalized for arriving late to a household, but receive no compensation for the additional time spent completing tasks once they arrive.
In both sectors, the boundary between paid and unpaid labor is continuously redrawn in ways that expand the amount of work while narrowing what is counted and compensated. Emerging subscription-based models reinforce this arrangement by requiring workers to pay a periodic fee in order to undertake work, thereby creating forms of economic lock-in that betray platforms’ promise of flexibility.
Workers’ accounts also trace how value is distributed across platform value chains. As one worker representative described, “A ten-dollar hourly rate billed to the client becomes two dollars in the worker’s hands.”
Workers’ accounts also trace how value is distributed across platform value chains. As one worker representative described, “A ten-dollar hourly rate billed to the client becomes two dollars in the worker’s hands.” Data workers described subcontracting structures in which a task billed at a significantly higher rate by a lead firm is reduced to a fraction of that value by the time it reaches the worker, with each intermediary layer extracting a share. Care workers similarly pointed to high platform commissions and opaque deductions that make it difficult to understand how their earnings are calculated.
Workers demand that the rate shown should be the rate received, to clearly show how value is distributed across these layers, and who is responsible for that distribution. They call for employer responsibility for all work-related costs, irrespective of classification; clear and predictable pay structures; recognition of, and compensation for, all working time, including waiting and commuting. Further, workers demand clearer allocation of responsibility across subcontracting chains, including forms of joint liability for lead firms and other intermediaries, to prevent the diffusion of accountability.
Moreover, it is tied to workers’ employment status—making platforms’ responsibility for work-related costs contingent on workers being classified as employees — even as platform business models actively misclassify workers.
The current draft Convention engages only partially with these dynamics. While it requires timely payments, it does not mandate transparent pay structures, regulate subcontracting-related wage erosion, or guarantee that work-related costs are borne by the platform. The restriction on charging such costs to workers appears only in the non-binding Recommendations. Moreover, it is tied to workers’ employment status—making platforms’ responsibility for work-related costs contingent on workers being classified as employees — even as platform business models actively misclassify workers. In this sense, classification itself operates as a bottleneck, with rights continuing to flow through formal status rather than attaching to the conditions under which work is performed.
Algorithmic control without meaningful transparency
If the externalization of costs defines how platform work is financed, algorithmic management defines how it is controlled. Across both data work and platform care work, workers described how their labor is governed via opaque processes and automated systems.
Data workers shared how task allocation is unpredictable, with no visibility into how work is distributed or why opportunities suddenly disappear. Similarly, workers do not know how their performance is assessed, what metrics are used, or how those assessments affect future work opportunities. Such decisions are made with little clarity on how automated processes and human oversight interact. Compounding these opacities, workers have no right to access the data generated through their labor, nor any ability to control how that data is used.
Platform care workers reported encountering similar mechanisms of control, although these were mediated by customer interactions through rating and feedback systems. Workers described being penalized or deactivated on the basis of complaints that are exaggerated, retaliatory, or simply false, with no meaningful opportunity to contest these decisions. Moreover, customers, who exercise significant control over working conditions, are not subjected to comparable forms of accountability. To counter this asymmetry, many workers demand that they be able to also publicly rate customers, with real consequences.
In both sectors, surveillance transcends mere task-performance to encompass the full scope of workers’ conduct.
In both sectors, surveillance transcends mere task-performance to encompass the full scope of workers’ conduct. Workers reported monitoring of output, behavior, and movement, including location-tracking and idle-time measurement that extended beyond the scope of work. Algorithmic systems function as infrastructures of evaluation, where workers continuously adapt their behaviour as they anticipate how access to work is determined, even as the metrics by which they are assessed remain unclear.
While the draft ILO Convention acknowledges platforms’ use of automated systems for work management, its protections are oriented towards individual decisions made via such systems, rather than their overall functioning. It requires platforms to inform workers of the use of automated systems, and in certain cases, to provide a human review for specific decisions. However, it fails to provide for meaningful transparency, withholding data from workers and the auditing authority from regulators. It also ignores the opaque ways in which automated systems and platforms’ organizational processes converge to exert control over workers.
In contrast, workers’ demands focus on more systemic transparency as they seek to assert agency over the digital infrastructures that govern their work.
In contrast, workers’ demands focus on more systemic transparency as they seek to assert agency over the digital infrastructures that govern their work. They call for democratic governance of data, visibility beyond the mere use of automated systems, and for how they concretely affect working conditions and access to work. They demand access and portability of personal and work-generated data, as well as the ability to influence how such systems operate. They also call for accountability mechanisms that extend beyond platforms’ internal processes. These demands locate algorithmic systems as central to the organization of work, and as sites that must be subject to oversight and contestation.
Transnational production without transnational accountability
Platform value chains stretch across continents. Meta, Microsoft, Google, and others contract their data-training needs to intermediaries like Sama (headquartered in San Francisco), which in turn subcontract to BPO (Business Process Outsourcing) firms in Kenya, Uganda, and elsewhere, where the actual work of data annotation and labelling is performed. This structuring of business through multiple nodes, offshoots, and intermediaries makes it challenging for workers to identify who ultimately controls their work and bears responsibility for its conditions. Worker-led initiatives, such as the collaboration between the African Content Moderators Union (ACMU) and Personaldata.io under the Data4Mods project, have begun the task of uncovering these outsourcing chains.
This structuring of business through multiple nodes, offshoots, and intermediaries makes it challenging for workers to identify who ultimately controls their work and bears responsibility for its conditions.
But even where these relationships can be traced, the challenge of enforcing responsibility remains. A platform worker in the Philippines, organizing with the BPO Industry Employees Network (BIEN), described attempting to invoke domestic labor protections: “We were told that even if the Philippines were to determine that the lead company was our employer, they wouldn’t be able to enforce laws because this is not where the company is headquartered.” The need for transnational regulation is clear, but the current draft of the ILO Convention does not meet the need. While allowing for bilateral or multilateral agreements on the regulation of platform work, it defaults to the laws of the country where the work is performed. Labor is organized across borders, while regulation remains nationally bounded.
Platform labor across sectors is often performed by migrants whose precarious legal status intensifies their vulnerability to exploitative working conditions.
Migration is central to the transnational organization of platform work. Platform labor across sectors is often performed by migrants whose precarious legal status intensifies their vulnerability to exploitative working conditions. In Europe, delivery work is disproportionately undertaken by South Asian migrants; in North America, app-based drivers are often migrants; in the Gulf and MENA regions, platforms play an increasing role in recruiting domestic workers from South and Southeast Asia; and across the world, migrants contribute a significant share of platform-mediated care work, including nannying and domestic services. In these contexts, migration regimes shape workers’ ability to refuse work, change employers, or seek redress. Workers pointed to how these arrangements operate in practice: domestic workers in MENA and the Gulf have reported deceptive recruitment processes followed by passport confiscation, while refugee data workers in East Africa described high wage deductions linked to social protection they are never able to access.
Despite this, the current draft of the ILO Convention does not meaningfully engage with the specific conditions of migrant labor. It does not address recruitment practices, the enforceability of contracts across jurisdictions, or the portability of benefits for migrant workers.
Exposure to harm without corresponding protections
The current draft of the ILO Convention also falls short in addressing the unique health and safety risks that platform work creates. Instead, the Blue Report backslides from the previous drafts of the Convention by removing obligations on platforms and placing responsibility primarily on governments.
For data workers — particularly content moderators who ‘clean’ and make safe the internet for all of its users — the need for tailored mental health protections is evident. Workers tasked with reviewing violent, sexual, and disturbing material reported sustained psychological strain, including anxiety, trauma, and symptoms consistent with post-traumatic stress. These harms are embedded in the labor process. Yet, the draft Convention does not treat mental health support as a core obligation.
For location-based platform workers, the risks are tied to how public infrastructure fails them as the city itself becomes their workplace.
For location-based platform workers, the risks are tied to how public infrastructure fails them as the city itself becomes their workplace. Workers described being forced by gamified algorithmic incentives to remain constantly available to provide services, often within tightly compressed time windows. They reported long-term health complications caused by prolonged periods without access to rest areas or sanitation facilities. Care workers, most of whom are women, pointed to the need for access to sanitary washrooms, especially as customers frequently deny them access to these facilities within the homes where they work.
The gaps in occupational health and safety are most evident in workers’ continued exposure to violence and harassment, including frequent instances of sexual harassment. Care and domestic workers, who operate within the ‘private’ sphere, reported heightened risks with few avenues for recourse. Platform beauticians in India described ‘emergency helplines’ being operated via AI chatbots, leaving workers in distress on hold for durations up to 30 minutes. Domestic workers in South Africa reported instances of workplace injury, such as being attacked by dogs in the households they worked in, without any healthcare support by the platform. While the ILO’s Domestic Work Convention (No. 189) recognizes households as workplaces subject to labor inspection, the draft Convention does not extend this logic.
An unclear regulatory horizon
The negotiations underway at the ILO will not simply extend labor protections into the platform economy. They will also shape how this new and growing form of work is understood and governed.
Across the discussions reflected here, workers describe arrangements in which the allocation of cost, control, and risk is misaligned with the distribution of responsibility. The draft engages elements of this reality through provisions on payment, automated systems, and health and safety. However, it does so in ways that leave underlying structures largely intact. Responsibility continues to be mediated through employment status, national jurisdiction, and the notions of an identifiable workplace, even where these categories are unsettled.
What the negotiations ultimately produce remains uncertain. Any standard will emerge from compromise. It remains to be seen whether worker demands for responsibility across subcontracting chains, meaningful accountability in algorithmic decision-making, and protection from the costs and harms of platform work will be translated into binding obligations.
It remains to be seen whether worker demands for responsibility across subcontracting chains, meaningful accountability in algorithmic decision-making, and protection from the costs and harms of platform work will be translated into binding obligations.
The implications will also unfold in national implementation, which will be shaped by uneven regulatory capacity and by firms’ ability to reorganize operations through subcontracting or withdrawal from particular markets.
Further, the Convention will not operate in isolation. Developments in labor organizing, data governance, competition policy, migration regimes, and corporate accountability are already shaping the contours of platform labor. The standard that emerges from the ILO will interact with these developments, contributing to a regulatory landscape that is still in formation.
