After years of sustained organizing and pressure by workers, unions, and allied movements across continents, the International Labour Organisation (ILO) agreed to establish binding labor standards for the platform economy. The decision marked the end of protracted negotiations between representatives from governments, workers’ unions, and employers’ organizations from 187 member-states at the 113th International Labour Conference (ILC) in Geneva in June 2025. While the bulk of the text is yet to be agreed upon, representatives finalised some crucial aspects shaping the standards—including their form, scope, and key definitions. In the coming year, following further negotiations on successive iterations, the ILO is expected to adopt a binding Convention and a non-binding Recommendation at the 2026 ILC, with substantive protections for platform workers across diverse legal and employment regimes.
In this context, this two-part article situates the global push for standards on platform work within the broader political economy of platform capitalism and the lived realities of its dispersed workforce. Informed by a series of interviews with workers, worker-organisers and platform economy researchers from geographies including Africa, South Asia, Asia-Pacific and Latin America, it attempts to unpack the global architecture of the platform economy, tracing its visible interfaces and invisible cross-border infrastructures, and suggests key frontiers for workers’ struggle for future-proofing the platform economy. In this Part A, Section I maps the scale, reach, and asymmetries of the platform economy, showing how North-South inequalities are hardwired into platform value chains and business models. It highlights how platforms outsource risk and extract value through complex layers of subcontracting, algorithmic management, and opaque control, rendering many workers— especially in the Global South—invisible and unprotected. From content moderators in Kenya to warehouse workers in Amazon’s global empire, the article explores how platform capitalism blurs traditional employment boundaries while reinforcing racialised and gendered labor hierarchies. Section II examines the resulting immiseration of platform workers: intensifying precarity, surveillance, and wage suppression, compounded by legal fictions of self-employment and the cross-border evasion of accountability. It foregrounds the extractive logics of digital rentiership and the structural limits of national regulation in confronting transnational platform power.
The forthcoming Part B will chart emergent strategies to reimagine labor futures: universalizing labor rights beyond employment status, reclaiming data and digital infrastructure, building worker-centric platforms, and challenging international trade regimes that entrench dominance by transnational digital corporations. Across these struggles, the article calls for systemic transformation grounded in cross-border solidarity, digital sovereignty, and democratic control over the future of work.
Part A: Extractive infrastructures–Platform capitalism and the new global division of labor
I. Beneath the interface: The visible and invisible architecture of the platform economy
The platform economy has expanded at a remarkable pace in just over a decade. As per the ILO, the number of digital labor platforms grew from 193 in 2010 to more than 1,070 in 2023. These include online platforms, which facilitate online work through a dispersed workforce, and location-based platforms, which connect workers and clients at specific locations. While the paucity of reliable numbers poses difficulties, the World Bank estimates that online platforms, by themselves, may engage up to 435 million active workers, i.e., up to 12.5% of the global workforce, a number which is likely to surpass one billion by 2030.
Despite this proliferation of platforms in the world of work, the architecture of the platform economy is far from even, and is shaped by persisting inequalities. Of the 545 platform-firms surveyed in 2023, 60% are based in the US and the EU. The North-South asymmetry is also reflected sharply in the global distribution of demand and supply of online work. More than 75% of the demand comes from high-income countries—primarily, the US, UK, Australia, Canada, Germany, and New Zealand. But the labor is performed disproportionately in the South—with India, Pakistan, and Bangladesh together contributing more than half of the supply. Even if one were to consider location-based platforms, a similar pattern emerges as one scratches the surface, with revenue generated from transactions in the South flowing predominantly (and largely untaxed) to firms headquartered in the US and China. While location-based platforms and online platforms together generated a global revenue of at least US$52 billion in 2019, 70% of this revenue was concentrated in these two countries. Notably, firms like Ola (India), Gojek (Indonesia), and Rappi (Colombia) have risen in Global South markets as homegrown challengers to global platform-behemoths. But despite their local origins, data from Tracxn reveals that they rely substantially on funding from actors in the US, China, or the EU.
Moreover, these figures represent only the tip of the iceberg, covering only relatively visible and well-documented forms of work, where platform-apps connect an individual worker on a consumer’s demand. The platform economy is, in fact, much more extensive, including a labyrinthine labor infrastructure and a host of less visible working arrangements entangled in cross-border value chains.
In India, ride-hailing apps Uber and Ola have been known to outsource management of certain taxis to fleet owners, which act as proxy employers. Drivers working under them face significantly different working conditions than those engaged directly by the app. The fleet companies control work allocation, impose additional rules, and sometimes operate their own surveillance systems. In Brazil, delivery platform iFood uses subcontracting firms to control workers’ schedules, often assigning shifts over private messaging apps and forcing workers to work for hours beyond the legal limit for employees.
Then there are workers like those servicing Amazon’s vast warehousing empire. These warehouse workers do not interact with customers on an app, but they perform essential logistical tasks: receiving, stowing, picking, and packing goods to fulfill orders initiated through Amazon’s platform. As Dharmendra Kumar of Janpahal, which works closely with the Amazon India Workers Association (AIWA), explains:
“Even though warehouse work isn’t ‘gig work’, it often resembles it. Many workers are hired part-time through apps, and there’s gamification of regular tasks. They face unrealistic targets and are constantly monitored through AI-based surveillance systems, through devices like scanners, with display boards showing real-time performance. If you fall behind, you get warnings from the ADAPT (Associate Development and Performance Tracker) system. After three warnings, you’re fired and blacklisted from all Amazon warehouses in India.”
Still deeper in the shadows are “ghostworkers”—those in the South who train AI systems and moderate content for platforms that we use every day, often through outsourcing firms. Meta’s now-infamous arrangement with SamaSource, a San Francisco-based firm, recruited Kenyan workers to screen content posted on Facebook and Instagram. Fasica Berhane Gebrekidan, a displaced Ethiopian journalist, recounts the misleading circumstances under which her ordeal with SamaSource began:
“In 2021, during the war in Ethiopia, I had to flee to Nairobi because of my Tigrayan descent. I had been working as a journalist for about five years in a state-run media outlet. I had no political affiliations or interest in politics—I never even wrote on political issues. But simply because of my name, which identified me as Tigrayan, I was targeted. I feared for my life. There were mass arrests happening in Addis Ababa at the time, and it was an incredibly stressful and dangerous situation. After arriving from Ethiopia to Nairobi, I heard about a job opportunity through a friend. He told me a company (SamaSource) was looking for people who could speak Amharic and Tigrinya. I sent in my CV and email, and about a week later, they contacted me for an interview. There were two rounds—one about my work experience and background, and another assessing whether I could psychologically handle the nature of the job. But at no point did they disclose that the job involved watching harmful content.”
Exposed to content depicting graphic violence and abuse, workers faced insurmountable performance targets to add to psychological trauma. When legal pressure mounted in Kenya, Meta switched contractors and moved operations to Ghana—where conditions are even worse.
Further, there are workers in “platformizing workplaces”—traditional work arrangements now governed by platform-style management tools. This is the silent creep of algorithmic control into retail, consulting, healthcare, and office work, where productivity is tracked not by a boss but by software. A recent employer survey in six OECD countries found that up to 90% of companies use some form of algorithmic management software. For instance, ActivTrak, an employee monitoring tool used by 9,000 companies, allows managers to access activity logs in real-time and alerts them with alarms when employees engage in undesirable behavior on their systems. When integrated with a customer-relationship management tool, it can also rank employees along metrics like “average productive hours per day”. This may not be platform work in the conventional sense, but it is the extension of the platform logic, where pervasive digital surveillance stands in for worker management.
These examples serve to highlight that the platform economy is not confined to the world of apps and gig workers. It is a vast, multi-layered system powered by digital infrastructures and dependent on precarious, fragmented labor. From taxi drivers in Sao Paulo to content moderators in Accra, from Amazon warehouse workers in Bengaluru to sales staff in financial corporations, these workers are bound together by the unrelenting pressures of pervasive data extraction, algorithmic subordination, and diminishing control over work.
II. Immiseration: The default platform worker predicament
A. “Flexploitation” and “digital rentiership” in the global race to the bottom
At the heart of the platform economy lies near-universal “flexploitation”, i.e., “the exploitation of workers under the guise of flexibility”. Classified under entrepreneurial labels such as “independent contractors”, “self-employed workers”, “service affiliates,” or “ partners”, workers are denied fundamental labor protections reserved for employees under labor law. This is despite an ILO study indicating that most workers across the South depend upon platform work as their main source of income. As Sangam Tripathy, National Advisor of the Indian Federation of App-based Transport Workers (IFAT), an affiliate of the International Alliance of App-based Transport Workers Union (IAATW), outlines:
“It’s clear that widespread unemployment is a major factor driving platform work in the region—whether in India, Bangladesh, Sri Lanka, or Nepal. After COVID, many people, especially those in low-skilled blue-collar jobs, lost their livelihoods. With few decent employment options available, platform work often became the only means of survival— something people could log into and start immediately.”
The denial of minimum wages, overtime pay, sick leave, health insurance, and protection from unfair dismissals is not incidental; it is central to how dominant platform-firms operate. James Farrar, former general secretary of the App Drivers & Couriers Union (ADCU) and a key litigant in the 2021 UK Supreme Court ruling that recognized Uber drivers as workers, explains:
“The first (reason) is the avoidance of tax, specifically employment-related taxes. The second is that they want complete variability in labor costs. They only want to pay for the work they believe is being done, with no obligation to cover downtime or other non-productive periods. And thirdly, it’s about network effects. For the business model to function, they need an excess supply of labor available at all times. That’s how they can offer quick ETAs and the kind of immediacy expected from tech-driven service delivery. But to make that economically viable, they can’t be paying for workers to wait around.”
The exclusion of workers from social protections also implies that the burdens of maintaining their health, livelihoods, and daily lives—what scholars term “social reproduction”—are fully borne by workers themselves.
This erosion of labor protections is accompanied by a deeper transformation in how value is generated and extracted in the platform economy. Strategies of “digital rentiership” allow platform-firms to profit not by producing goods or services themselves, but by controlling digital infrastructures—apps, algorithms, data—that participants in the platform economy must use to interact. Backed by venture capital to rapidly leverage network effects, lock in users and drive out (or subsume) competitors, often even at the cost of productivity and service quality, these firms collect monopoly rent—taking a cut from transactions, while also continuously capturing vast amounts of data.
Besides fueling further expansion of their services in new markets, data also powers the algorithmic systems that platform-firms use to dynamically and opaquely control the labor process. Farrar underscores the implications of this algorithmic opacity for workers:
“It’s a basic principle everywhere that workers should have access to the particulars of their employment (including) actual pay and working conditions. But how can a gig worker have that? You show up to work without knowing if you’ll actually get any work, and even if you do, you don’t know how much you’ll be paid for it.”
In strained labor markets with an oversupply of workers, algorithmic wage-setting mechanisms compel the most vulnerable workers to constantly bid against each other, pushing down wages further. In the words of Sofia Scassera, a researcher on the digital economy with the Transnational Institute (TNI) and an advisor to international trade unions:
“The aim of managing workers this way is to create information asymmetries that push workers to compete with one another—being more available, taking on more tasks, and overworking themselves to be seen as the ‘best’ and earn better benefits. It’s a system designed to extract more from labor through competition and uncertainty.”
On online freelancing platforms such as Amazon Mechanical Turk (AMT), where “Turkers” around the world compete for “Human Intelligence Tasks”, this race-to-the-bottom assumes a global dimension. For example, a study by David B Martin and others concluded that Indian AMT workers are likely to accept lower wages, given that all workers on AMT are paid in US dollars irrespective of their location. Research by Anita Gurumurthy and others suggests that Amazon’s algorithmic systems know this only too well, specifically targeting lower-paying tasks for Indian workers.
B. Gendered and racialized hierarchies
The relatively low requirements of capital investments and skills in the platform economy as a compelling proposition for workers otherwise at the margins of labor markets. Dharmendra observes in the Indian context:
“In the absence of formal, secure jobs, gig work becomes a kind of refuge. Many people don’t have the skills to access salaried employment or the capital to start a business, but if they can get basic driving skills and a bike, they can enter the gig economy. Financing a vehicle is also easier now because of the rise of a “buy now, pay later” culture that financial institutions have normalized. Agricultural growth has been stagnant for decades, which has led to large-scale distress migration to cities. But in urban areas, most migrants only find informal or manual jobs—loading, vending, or driving cycle-rickshaws or e-rickshaws. Compared to (these other options), gig work can seem more attractive, especially for those who are educated, can own a motorbike, and don’t want to take up traditional manual labor.”
The account of Ebony Ford, a taxi driver and Permanent Secretary of the Tanzania Online Drivers’ Association, aligns with these grim economic realities:
“The motivation at that particular time was that after schooling, I had no job, and there was nothing around. I was running a small business that enabled me to have a car. So, I decided to take my car into the business and drive in Dar Es Salaam.”
At the same time, in its interactions with deeply stratified societies and economies, the platform economy reproduces—and often intensifies—existing hierarchies.
The ride-hailing and delivery sectors, which usually drive policy action, are heavily male-dominated. Women encounter similar hindrances to meaningful participation in these sectors as they do outside the platform economy—limited access to necessary equipment (smartphones and vehicles) and finance, and safety concerns, reinforced by gender norms. Muhammad Amir Anwar’s research with drivers in South Africa, for example, reveals that while a majority of drivers prefer driving at night due to higher demand and less traffic, safety concerns lead women drivers to avoid such trips.
Women are indeed better represented in care, beauty, and domestic platform work—areas long marked by feminized, underpaid labor. Yet, having to bear the double burden of paid work and unpaid care responsibilities consigns women to fear and less rewarding opportunities than their male counterparts. As one woman worker from Africa observes in this consultation by IT for Change:
“Not only are women having to combine productive and reproductive labor in ways that men aren’t, but they’re actually working potentially longer hours than men in the process.”
The lack of gender-responsive social protection only compounds this inequity, leaving women disproportionately exposed to economic shocks. As another platform worker laments in the same report:
“The lack of social protections is devastating. There is no paid maternity leave, sick leave, or parental benefits. If a worker’s children fall ill, she loses her income.”
Just as platform work deepens patriarchal labor divisions, it also entrenches racialized stratification. Across metropolitan centers, location-based platform work is carried out disproportionately by migrants, with a large section from marginalized racial and ethnic identities. Qualitative research by Van Doorn and others, spanning cities like Amsterdam, Bangalore, Cape Town, and New York, suggests that it is the constant influx of precarious migrant labor seen as disposable, which makes platform-firms’ business models viable.
In Berlin, Orth notes that platforms like Deliveroo recruit migrants with limited German language skills and insecure residency status, such as student and working holiday visas, which facilitate unstable employment relationships. As a Chilean courier quoted in Altenried’s research reflects: “…this is the only option that the immigrants (in Berlin), or people from Chile, or people from India have. So, even though the work conditions are shit, (…), I was really happy with it. And as long as I didn’t get hit by a car, everything was going to be okay”.
A similar story plays out in his native Chile, where migrants constitute 60% of all workers engaged on location-based platforms, stuck in low-wage, dead-end jobs with no real path to mobility. In a study by Bonhomme and Muldoon, a Venezuelan courier navigating the daily grind of app-based delivery in Chile, captures this sense of stagnation: “Projections don’t exist here (…) you’re a slave to the application (…) you don’t go anywhere from here, where would you go?… senior delivery man? No, that’s not going to happen. They don’t do anything for you. Here you’re a delivery boy, you do that, and that’s it. And God willing, nothing bad will happen to you.”
Beneath the veneer of inclusion, then, the platform economy funnels the most vulnerable sections of the population into the most precarious forms of work and keeps them there, with little opportunity to move up or out.
C. Barriers to collective organizing
If the platform economy is predicated on disempowerment, perhaps its most enduring achievement has been in making collective resistance seem gruelling. Platform-firms fragment the workforce, atomize workers, and deploy advanced forms of digital surveillance to defuse collective action before it takes shape. Again, the (mis)classification of workers aids this, depriving them of collective bargaining guarantees, including the right to unionize— and in certain jurisdictions, even recasting their attempts to collectively demand better working conditions as collusive.
Simultaneously, the geographical dispersion of workers, the instability of the workforce, and the lack of common physical spaces for collectivization serve to reinforce the construction of the worker as a self-dependent entrepreneur. Ebony recounts how drivers in Dar es Salaam navigated logistical challenges to expand the membership of the Tanzania Online Drivers’ Association:
“We started a WhatsApp group where, whenever we met someone driving for a platform, we’d ask them to share numbers of other drivers they knew. We used the group to explain the issues we were facing and discuss what we could achieve if we had an association. Sometimes we even put up a QR code to join the WhatsApp group at nearby platform offices. People would scan it and join directly. It wasn’t easy, but we put in a lot of effort.”
Contemporary digital surveillance extends the capacity of platform firms to monitor and crack down workers’ attempts at organizing. For example, Amazon-owned Whole Foods collects data on union presence, local economic conditions, and union membership to map potential sites of union activity. In Berlin, the grocery platform HelloFresh actively flagged and removed pro-union posts by workers on social media.
In cases where workers succeed in making their collective voices heard in public, reprisals are swift and strategically punitive, as Dharmendra highlights:
“In Patna, India, food delivery workers staged a mass logout from platforms like Zomato. In response, companies used two main tactics. First, they offered time-bound incentives to non-participating workers and temporarily relaxed shift rules, allowing workers to log in after 6 p.m. to create the illusion of normalcy. Second, they retaliated—some workers had their IDs deactivated, with one app message labeling a participant a “strike enabler.” This shows how in the platform economy, the right to strike is not only denied but in effect criminalized.”
D. Cross-border evasion and the crisis of enforcement
Even as national regulators grapple with platform-firms’ constantly evolving business models and their effects on local economies, these firms engage in “regulatory entrepreneurship”, bending the rules as they are being written, as a concerted business strategy. A 2020 campaign in California provides a stark illustration: Uber, DoorDash, Lyft, Instacard, and Postmates together spent more than $200 million to mobilize consumers as well as workers against a state law, AB-5, which required them to treat their workers as employees. In promoting the ballot initiative that would exempt them from the law, they targeted residents with coercive in-app notifications and misleading ads issued in the name of non-existent organizations like “Progressive Voter Guide”. In Brazil, during mass delivery worker strikes in 2020, iFood anonymously contracted marketing agencies to promote messages such as “without a boss and minimum wage, delivery workers have more freedom and earnings”.
As Scassera explains, such instances reveal that building robust public accountability mechanisms is less a regulatory challenge and more a fundamentally political struggle:
“Perhaps the biggest challenge is the political struggle—how transnational corporations are viewed by public and political leaders. Opposing powerful platforms often comes at a political cost, with leaders risking being seen as anti-development or anti-technology. Even progressive governments hesitate to challenge Big Tech, preferring the narrative that equates technological progress with deregulation.”
Even where laws are enacted, eliciting meaningful responses from transnational platform-firms before courts can be a challenge. In well-documented cases in the UK and South Africa, Uber invoked the corporate autonomy of its Dutch holding company to dodge the court’s jurisdiction or delay compliance with rulings. In Spain, Deliveroo outright exited the market after the government passed the Riders’ Law, which recognized app-based delivery riders as employees. In Kenya, Meta’s defense in the 2023 lawsuit by content moderators included the claim that Kenyan courts lacked jurisdiction over the US company. Such evasion is indicative of a broader crisis in cross-border enforcement, which is exacerbated by the ability of digital enterprises to carry out their operations without physical assets.
Moreover, as Kate Lappin, Asia-Pacific Secretary of Public Services International warns:
“Amid ongoing trade wars, there seems to be momentum toward restrictive trade agreements. One major issue is that such agreements prohibit states from imposing requirements of local presence on corporations. Without a local presence in the countries where they operate, you can’t have an effective tripartite system—the foundation of international and most national labor laws. There’s no way to bargain or enforce accountability. If a company refuses to pay or closes down, there’s no legal remedy if they hold no assets locally.”
E. Algorithmic opacity and data deficits
Another significant yet underregulated dimension of the platform economy is the opaque data infrastructure that underpins algorithmic management, surveillance, and exploitation of workers. The information asymmetry not only further undercuts workers’ capacity to negotiate their working conditions, but also obscures the scale of wage theft that platform-firms engage in. In a piercing 2024 report, the aforementioned WIE estimates that data currently held by ride-hailing and delivery platform-firms in London masks as much as £1.9 billion of unpaid wages.
Admittedly, the EU’s General Data Protection Regulation and similar privacy laws in other jurisdictions offer some tools to unearth opaque algorithmic decision-making. But such laws’ reliance on due process overlooks the fundamentally unequal power dynamics between workers and platform firms. Farrar, who now leads Worker Info Exchange (WIE), a non-profit that advocates for workers’ informational rights, recognizes these limitations:
“Firstly, it falls on individual workers to make complaints or pursue legal action themselves. But that’s often not a realistic option—especially for very precarious workers who don’t have the time, resources, or security to do that. Secondly, without a basic level of transparency to begin with, how would you even know something’s wrong? How do you know what to ask for, or whether your employer is acting within the law? You wouldn’t. If everything is happening below the surface, how are you supposed to know what you don’t know?”
Further, legal scholar Veena Dubal points out that laws on personal data do not apply to insights derived from aggregated “depersonalized” data, which are key to the functioning of algorithmic work management systems. Dubal expresses similar apprehensions regarding the EU’s new framework on AI, and its specialized directive on platform work, noting that “transparency and information-sharing are not immediately useful in the context of a workplace in which digital systems are constantly changing.”
Moreover, while data laws can be, and are indeed being, creatively leveraged to contest the uses of data in worker surveillance and algorithmic control in the platform economy, there is no comprehensive framework that recognizes the economic value of workplace data for workers or enshrines collective worker data rights. Farrar emphasizes that without a collective lens, both the harms of algorithmic management and the means to resist them remain obscured:
“You can’t really understand the effects of something like discrimination without looking at it collectively. You won’t see the full picture just through individual experiences. But beyond that, collective action is how we address the power imbalance between workers and employers—especially when workers are treated as individual contractors. That’s the whole point of collective action in the workplace: it allows workers to pool their power and engage in a more equal, balanced dialogue with employers. This has been the foundation of industrial relations for at least the last hundred years. And as we move into the digital economy, we must approach digital industrial relations in the same way: by finding ways to collectivize.”
Building on the analysis of structural asymmetries, worker immiseration, and regulatory erosion in this Part A, Part B will outline existing initiatives and examine how the labor movement can confront confronting platform power through efforts to expand labor rights, reclaim control over data and digital infrastructures, challenge the international trade regime and redesign the platform economy in the interests of workers.
