The gig platform's founding claim was liberation: no boss, no schedule, be your own business. The claim was not entirely false, which is what makes the arrangement interesting. A rideshare driver genuinely has no manager watching her, no meetings, no performance review with a human on the other side. What she has instead is a landlord.
Consider the structure rather than the branding. The driver supplies the capital (car, phone, fuel, insurance) and all of the labor. The platform supplies access to the field where demand grows, and takes a share of every harvest that crosses it. It sets the terms of tenancy unilaterally and can change them without negotiation. It watches everything, prices everything, and can evict (the industry word is deactivation) through a process with no court, no confrontation with an accuser, and often no stated reason a human will defend. The relationship is not employer and employee, and it is not seller and customer. It is lord and tenant farmer, rebuilt in software.
Management by algorithm, jurisdiction by design
The scale of the arrangement is now considerable: the International Labour Organization counted at least 653 active digital labour platforms globally as of October 2025, while noting that conventional labor statistics still under-capture platform work (ILO, 2025). And the research literature has converged on a name for how these platforms govern: algorithmic management. A 2026 systematic review of 103 empirical studies documents platforms using algorithms to run the classic functions of a boss (setting pay, scheduling, monitoring, rating performance) across companies from Uber and Meituan to Fiverr (Frontiers in Sociology, 2026).
The word management undersells what is happening, though. A manager operates inside an employment relationship that law constrains; the algorithm operates a jurisdiction. It writes the rules, detects violations, imposes penalties and hears the appeal, all in one opaque process. The driver rating that quietly determines a worker's access to good jobs functions like standing in a manor court: consequential, unappealable, and maintained by the same power it judges. This is the feature that employment-law remedies keep sliding off. Reclassifying workers as employees addresses wages and benefits, but the deeper asymmetry (one side writes and enforces all rules of the shared world) survives reclassification untouched.
The tell: coordination without lords is possible
It would be convenient for platforms if this architecture were simply what coordination at scale requires. The historical record of the internet itself says otherwise. Yochai Benkler documented, in his analysis of Linux and open-source production, a third mode of organizing beside firm and market: commons-based peer production, in which thousands of contributors coordinate complex work without managerial hierarchy or price signals, governed by participation and merit rather than employment authority (Benkler, 2002). The technology that runs the platforms was substantially built this way. Distributed coordination without a lord is not utopian speculation; it is the platforms' own supply chain.
So the tenancy structure is a choice, and the choice has a logic: the toll. A platform that merely matched supply and demand, portably and neutrally, could not sustain its take rate; a worker who could carry her reputation and client relationships to a rival could bargain. The moat is the non-portability: the rating, the history, the client relationships all live on the lord's land and die at the border. Serfdom's economic core was never the work; it was that leaving meant starting from nothing.
What a free tenancy would look like
The feudal diagnosis implies remedies that differ from the usual employment debate. Portability: reputation and work history that belong to the worker and travel with her, the equivalent of being able to take your tools. Interoperability and multi-homing: the real freedom to farm several fields at once, undegraded by punishment mechanics. External jurisdiction: deactivation and disputes decided by processes not owned by one party. And transparent tenancy terms: the algorithm's pricing and allocation rules stated plainly enough to be contested.
None of this requires romanticizing the pre-platform world of dispatchers and taxi medallions, which had lords of its own. The point is narrower. The gig economy answered a real desire (autonomy, flexible hours, work without a boss's gaze) with a structure that delivers the feeling of independence while re-implementing dependence one layer down. The driver has no manager and less power than an employee; no schedule and less security than a tenant with a lease. The medieval serf, it is worth remembering, was also technically not a slave and technically attached to no master, merely to the land. The gig worker is attached to the app. The legal form is novel. The relationship is very old.
Sources
- ILO (2025, reported). Count of digital labour platforms. The ILO identified at least 653 active digital labour platforms globally as of October 2025, noting official statistics under-capture platform work. https://www.frontiersin.org/journals/sociology/articles/10.3389/fsoc.2026.1743445/full
- Frontiers in Sociology (2026). Algorithmic management in the global gig economy: an interdisciplinary systematic literature review. Review of 103 empirical studies documenting platform control of pay-setting, scheduling, monitoring and performance rating. https://www.frontiersin.org/journals/sociology/articles/10.3389/fsoc.2026.1743445/full
- Benkler (2002). Coase's Penguin, or, Linux and The Nature of the Firm (Yale Law Journal 112). Documents commons-based peer production: complex coordination without managerial hierarchy or price, governed by participation and merit. https://yalelawjournal.org/article/coases-penguin-or-linux-and-the-nature-of-the-firm