Digital labour platforms are rapidly transforming the world of work. Many governments in the global south have welcomed platforms that bring businesses, workers and consumers together, creating opportunities to reduce unemployment. But what happens when jobs are turned into tasks? Workers are increasingly managed by apps that allocate these tasks and monitor their performance. Platform workers face low pay, poor working conditions – and algorithmic surveillance.
There are two broad types of digital labour platforms. The first are “location-based”. They mediate tasks carried out in person at particular locations: delivering food, taxi services, domestic work and care services. The second are “online web-based”. They mediate tasks such as data categorisation and translation and editing services that can, in theory, be performed anywhere via the internet and remotely.
While digital labour platforms create new opportunities for income, they also threaten to extend informality into new sectors through “algorithmic insecurity”.
Platform work is often seen to create opportunities that better suit workers’ skills, interests and schedules. This increased flexibility is said to provide workers, particularly women who have to balance paid and unpaid care work, with income opportunities and greater autonomy over their work. However, a growing body of evidence shows that the flexibility of many of these jobs comes at a cost to workers’ economic security and control over the work process.
Many people are drawn to platform work because they aspire for the greater autonomy and control that comes from “not having a boss”. Yet, as Brazilian historian Lucas Santos shows in his research with food delivery workers in São Paulo, Brazil, riders soon realise that the “feeling of freedom” they associate with “working for an app” is more aspiration than reality.
Protesting food delivery workers in São Paulo, Brazil (Getty Image)
Indian sociologist Gayatri Nair’s research shows that many gig workers in India are attracted to the “initial façade of formality and respectability”. Many women workers in India’s beauty sector see “digital mediation” via an app as creating a “form of respectability” that was not there before. But, Nair shows, earnings are meagre and many workers lack autonomy and control over their work. Digital labour platforms, she argues, are creating new forms of “algorithmic insecurity” in the form of surveillance, ratings, and arbitrary changes to remuneration.
The reality of algorithmic control
Digital labour platforms often deny any employment relation. They insist that they act simply as a mediator between businesses, workers and consumers. What distinguishes them from a traditional labour market intermediary is that what they mediate is a single task, activity or service rather than a job in the traditional sense. And though platforms are never fully autonomous, they manage workers and tasks using algorithms (computer programmed procedures). In so doing they minimise human interaction.
This process of automating work processes and the coordination of tasks is often referred to as “algorithmic management” or “algorithmic control”.
For example, food delivery workers in São Paulo and Johannesburg, South Africa receive automated instructions on their mobile phones about where to collect and drop off the food, and the route via GPS map they should take. South African sociologists Eddie Webster and Fikile Masikane show that digital labour platforms not only assign tasks: they also assess performance and determine pay. They have the power to deactivate or disconnect a worker from the platform, unilaterally terminating employment.
Algorithmic management is invisible and inaccessible. It also controls all aspects of the work process. As Colombian sociologist Derly Yohanna Sanchez points out, this control extends to more than just the “supply and demand” for digital work. Platforms also control aspects of the work that are, she says, “external to the digital market such as personal information and human resources”.
This can produce anti-competitive or discriminatory outcomes, or both. For example, platforms give customers, as Nair put it, “full power to rate and review” workers. This algorithmic “disciplining” of workers via customer ratings has minimised transparency and accountability. It has also created perverse “information asymmetries”, as noted by communications scholar Sai Amulya Kommarraju. These make it nearly impossible for workers to contest customer ratings or challenge the forms of algorithmic control they are subjected to.
Although algorithmic control appears insurmountable, there is some evidence that workers are beginning to push back and fight for greater transparency and improved worker conditions.
In December 2020, Uber drivers in Johannesburg, South Africa launched a protest by disabling the Uber app and not accepting requests for rides. Among the drivers’ complaints were the obscure way in which their accounts were blocked by Uber and the inequitable way in which the fees earned by drivers were unilaterally decided and implemented by Uber.
If digital labour platforms are here to stay it is an imperative to build platforms and algorithms that prioritise the interests and needs of workers, not only business. Giving workers the right to access and have control over their data is, International Labour Organisation economist Uma Rani argues, one of the most important ways to “empower worker control over algorithmic management”. This will require both regulation and pressure from below, by consumers and workers
Article authors: Hannah J. Dawson, Senior Researcher, Southern Centre for Inequality Studies, University of the Witwatersrand; Ruth Castel-Branco, Research Manager, University of the Witwatersrand