Digital labor platforms like Uber, Gojek in Indonesia, or Ola in India have moved from the periphery of labor markets to the mainstream. As of late 2025, at least 653 active digital labor platforms operate globally across sectors including ride-hailing, delivery, freelance services, healthcare, and domestic work. The number of global online gig workers ranges between 154 million and 435 million, representing 4.4% to 12.5% of the global labor force. Even these figures likely understate the true scale, given persistent global measurement gaps.
Given their scale, it is important that governments and researchers be able to study digital labor platforms. But what information is available and who gets to access it? My research with colleagues at Utrecht University shows that platforms are highly selective about what data they share and with whom. That shapes what the public can know about them and thus what policy can do about them.
Digital labor platforms in the economy
In Brazil, recent national data indicate that 1.6 million people now work as platform-based drivers or delivery workers, nearly 2% of the country’s workforce. According to the Brazilian Institute of Geography and Statistics (IBGE), the number of platform-based workers grew 25.4% between 2022 and 2024, and their average earnings are higher than those of non-platform workers.
Bonina et al. (2021) reflect on the role that digital platforms can play in development. One of the positive roles the authors highlight is precisely that related to job creation in the service sector (Gojek is the example cited in the article), which can increase labor income and formalize the labor market.
India offers a telling example. Ride-hailing platforms such as Uber and Ola have been integrated into broader narratives about employment generation, smart cities, and digital entrepreneurship. State governments have experimented with regulatory partnerships and digital integration initiatives, framing platforms as instruments of modernization and job creation (for examples, see Rodrik & Sandhu (2024) and Rodrik (2025)).
At the same time, Bonina et al. note that the lack of labor legislation can lead to excessive exploitation of workers by these platforms in the service sector. In addition to the need for legislation to protect workers on these on-demand platforms, there is another major issue that must be addressed as platform work is integrated into development policies: who controls what we can empirically know about platform labor markets?
Data and epistemic selectivity
Governments often rely on platform-reported statistics to estimate driver numbers, earnings levels, and economic impact. Detailed data on net income after costs, algorithmic allocation systems, and commission structures remain largely proprietary. When development strategies rely on private digital infrastructures, informational dependency becomes a constraint.
In a recent article in Platforms & Society, together with scholars from Utrecht University, we have examined 397 empirical academic studies covering 57 labor platforms operating in Europe. Our central finding is striking: only 7% of empirical studies rely on platform-provided data.
Most research on platform labor, particularly on earnings, discrimination, algorithmic management, and working conditions, depends on interviews, surveys, ethnography, or scraping. When platforms do share data, access tends to concentrate around topics such as consumer welfare and sustainability. We call this pattern epistemic selectivity: favoring specific research agendas through selective data access.
This matters for policymakers. When millions depend on platform income and when average earnings in some contexts appear attractive relative to alternatives, public policy requires rigorous, replicable, independent evidence. If access to detailed labor data remains privately controlled, then governments, researchers, and labor organizations operate with partial visibility.
Our study identifies three models of platform data sharing: open aggregated portals; API-based access under platform-defined rules; and bilateral agreements, often involving elite institutions and non-replicable datasets. The most detailed datasets are typically governed by contractual arrangements and rarely publicly replicable. Such arrangements are frequently concentrated in well-resourced institutions, reinforcing global asymmetries in research capacity.
For countries integrating platform work into employment or urban development strategies, there are two risks that need to be avoided: economic dependence combined with informational dependence.
Takeaways
As we trying to design institutions capable of delivering good jobs and shared prosperity in the twenty-first century, platform labor is already part of that reality. With hundreds of platforms operating globally and millions of workers relying on them for income, including a rapidly growing share in countries like India and Brazil, the question is no longer whether platforms are inherently good or bad.
The real issue is whether democratic institutions have the informational capacity to govern labor markets increasingly mediated by private digital infrastructures. If platforms shape employment trajectories, urban mobility systems, and income generation, then researchers and public authorities must have systematic and transparent access to the data needed to assess those impacts. Without that capacity, labor governance risks being shaped by selectively curated evidence rather than robust public knowledge.
Read the paper: "The Epistemic Selectivity of Online Platforms: The Case of Data Sharing for Academic Research on Labour Platforms" by Victo Silva, Iryna Susha, Jarno Hoekman, Koen Frenken
