Data rights are labor rights, especially when it comes to the platforms of the gig economy. Leveraging data for the collective good is essential for the future of work and internet health.
When driving for Uber one night in London in 2015, James Farrar was assaulted by a passenger. What began as an uneventful Friday, ended with aggression that spilled out of his car and onto the roadway. It was a jarring event, but Farrar assumed that Uber would quickly identify the aggressors and report them to the police. Instead, there was silence for weeks, and he saw it as a sign of disrespect for him as a driver for the company.
It was among the first of many instances where Farrar felt Uber held data that intimately concerned him, although he could not access it. It caused him to pore over his contract, in which he noticed the emphasis placed on drivers being self-employed as opposed to employees. Farrar bristles at this: “If I’m my own boss, running my own business, and you’re just my agent — how come I can’t know who my own customer is?”
He describes the experience as the catalyst that got him engaged in campaigning for labor rights for app drivers. Together with another former driver, Yaseen Aslam, he later founded the App Drivers & Couriers Union (ADCU). After six years of legal battles, Farrar and Aslam currently await a landmark ruling by the United Kingdom’s Supreme Court on whether app drivers in fact are employees, and therefore entitled to rights like minimum wage and holiday pay.
Gig work is everywhere
An estimated 50 million gig workers worldwide toil within ecosystems created by online platforms such as Uber, Ola Cabs, iFood, Grab, Helpling, and dozens of others. It is a global phenomenon that contributes to the commodification of labor in the context of limited data rights for workers worldwide. And so, expanding data rights and protections would have ramifications for the future of work as well as for internet health.
“A majority of gig workers are in the Global South,” says Funda Ustek-Spilda, a researcher and project manager at Fairwork, an international research community studying the global platform economy. She notes that the risks associated with gig work disproportionately affect systematically vulnerable communities everywhere. In Europe and North America, for instance, gig workers are more likely to be people of color.
Powered by individual and aggregated data, the automated systems employed by platforms play a huge role in determining who is offered work at what price. At the same time, it is a business model that tends to rely on an oversupply of labor — a convenient system for consumers who want cheap, immediate service, but often leaves workers idling, underpaid, and at the mercy of opaque algorithms.
Digital rights are labor rights
In his own legal exchanges with Uber, Farrar discovered several details about the data they collected about him and what they inferred about him as a driver.
For instance, he learned that Uber maintained a secret profile that included electronic performance tags such as ‘missed eta’ or ‘negative attitude’ — tags Farrar attained for refusing work he believed would be unprofitable or for requesting reimbursement of an airport parking fee. Farrar was never informed of this nor does he know how the tags were processed. But he suspects performance factors play into work allocation systems.
“Drivers are led to believe that they are working in a completely open market, but if work is being throttled for some people due to performance factors, they deserve to know and have a right to appeal,” he says.
For individuals and collectives, gaining access to more personal data has the potential to reveal the inner workings of secretive processes. It demonstrates just how tightly bound digital rights now are to labor rights.
In response, a gradually emerging tenet of labor organizing is centered on unlocking power from data.
Christina J. Colclough of The Why Not Lab is an advocate for global labor rights who has dedicated years to highlighting why labor unions must see data rights and governance as an urgent priority. She sees digitization and worker surveillance increasing in numerous sectors, beyond the gig economy, and in ways that have only accelerated since the COVID-19 pandemic. “The tech world’s super surveillance is exploding, and the power asymmetries in the labor market are growing,” says Colclough. “Honestly, I would say that this is a situation where unless organized labor begins to respond, it will soon be too late. But the majority of unions are not engaging on this yet,” she says.
The picture is complicated by the fact that a lot of digital technology in the workplace is proprietary and developed by third-parties. The management of a company deploying such technologies may have limited rights to alter them, let alone know how to govern them. Colclough thinks unions should be thinking of ways to co-govern algorithmic systems and to negotiate for much stronger collective data rights for workers across what she calls the “data life cycle”. At every stage, from collection to analysis to storage, and potential transfer to third parties, there are rights to be negotiated, she says.
Data under lock and key
Requesting, obtaining and organizing data is more difficult in practice than describing it in theory. Gig work platforms have few incentives to share data they see as central to their business. Worse, when it comes to estimating things like hourly wages, companies have an interest in inflating numbers and obfuscating independent research.
“The issue, philosophically, is that these apps collect all of the information that we would ever need to know, but the only way to understand it from a research or worker-organizing perspective is still to gather the information yourself,” says Dan Calacci, a PhD student at the Human Dynamics group at the MIT Media Lab.
Calacci has contributed to a number of innovative technical projects to enable workers to collect useful data themselves. One is WeClock, an open source app created by partners including the UNI Global Union’s Young Workers’ Lab (led by Colclough and Jonnie Penn) together with Guardian Project and OkThanks. The app enables workers to “quantify” their labor to document their own work hours, travel distances, and pay rates, for collective action or evidence of productivity. The app keeps all data local and gives workers control over who it is shared with.
Another app, created by Calacci with the labor activism group, Coworker.org, is an SMS-bot called the Shopper Transparency Calculator. When Shipt, a large on-demand shopping app in the United States changed their pay structure from a flat rate to a blackbox algorithmic calculation, their “shoppers” could not estimate how it would affect wages. In October 2020, Calacci published a study based on thousands of screenshots from 213 workers who used the app showing that 40% of shoppers had their pay cut. As they protested, the results of the study were disputed by Shipt without further clarification.
Data futures in flux
Technically, there are challenges of enabling workers across many different gig platform apps (often used simultaneously) to better understand how their wages are calculated across different mobile devices, but Calacci says one of the biggest challenges is to develop models for data governance that are easily applicable in practice and appropriately sensitive to the risks and responsibilities of collecting sensitive personal data.
“Projects can say that they are concerned with privacy but then still go on to collect as much data as possible, even though it might not be in the best interests of workers,” he says.
In 2020, research by Mozilla’s Data Futures Lab concluded that there generally still isn’t clarity around how new ideas for data governance (like data trusts, data commons, data collaboratives, and more) will work in practice. Nor is it clear how it could be easier for grassroots initiatives in different jurisdictions to organize and manage data responsibly. To complicate matters, terms are often used interchangeably with different definitions.
In order to shed light on the questions that unions need to ask about data, Prospect in the United Kingdom launched a quiz developed by Keith Porcaro called Lighthouse. It is designed to help unions think about how to use and handle data (or not) with checklist items like, “Each of our digital assets have someone responsible for managing and safeguarding them”.
Colclough says most unions do not have data analysts on staff, but that she sees new organizations innovating on “collectiveness” in data governance.
Driver’s Seat Cooperative is one such organization. They enable app drivers in the U.S. to pool together data about their movements and earnings. “We’re ensuring that a broader set of stakeholders have access to the data,” says Hays Witt, co-founder and CEO of the cooperative, which is owned and governed by its members. “And we use this to give drivers insights into where and when the best place to work is, and which platforms pay better.” At the same time, Driver’s Seat also markets mobility insights to city planners and transportation agencies, sharing dividends of earnings with drivers.
Fights for fairness
Since the pandemic, collective action like strikes and unions are growing more frequent. In the spring and summer of 2020, gig workers in Brazil protested by the thousands, crowding into downtown São Paulo and demanding better pay, working conditions, and health safety measures from platforms like Uber and iFood.
“They are making us work weekends, every day, or we face the risk of getting blocked,” one iFood worker told Al Jazeera in July 2020. Similar protests spread across the continent, springing up in Mexico, Chile, and elsewhere, culminating in a movement dubbed “Breque dos Apps”. The strikes and protests brought media attention to gig workers, and galvanized support for a bill in the Brazilian parliament in July 2020 that would guarantee an hourly rate and paid vacation.
The central question of how internet and app-based gig work is classified has been hotly contested in different countries, with different outcomes. In California, leading up to elections in November 2020, Uber, Lyft, Doordash, and Instacart reportedly poured $200 million USD into a successful campaign for a ballot initiative (Proposition 22) that now exempts their gig workers only from certain benefits and protections. In September 2020, Spain’s Supreme Court ruled that drivers for the food delivery app Glovo, based in Barcelona, were employees not freelancers.
Regardless of employment classification, Fairwork argues in their assessments of gig platforms in different countries, that there are principles like fair pay, fair contracts, and fair management that all should adhere to. Some do better than others.
“One of our objectives as a union is to take control of our personal data,” says Farrar about the ADCU. In Europe, a tactic has been to support drivers in filing for access to personal information by leveraging provisions of the European Union’s General Data Protection Regulation (GDPR) that guarantees access to personal data held by companies. Even more leverage could be on the horizon, as improving protections for people working in the platform economy has been announced as a new key European Union initiative for 2021.
Longterm, the ADCU is working with Worker Info Exchange to collect and pool app-driver data in order to gain collective power in dealings with the companies [Mozilla is supporting this initiative with a grant through its Data Futures Lab]. But according to Worker Info Exchange, both Uber and another major platform, Ola Cabs, have deliberately frustrated data access rights through “lengthy and convoluted request processes often resulting in little or no data being provided”.
In response, the ADCU and Worker Info Exchange brought a lawsuit in July 2020 on behalf of 13 drivers in the United Kingdom against both Ola Cabs and Uber in The Netherlands. They requested access to “secret worker performance profiles” and “fraud probability scores” that neither company wishes to disclose. Uber has publicly countered that they have shared everything drivers are entitled to and would “never share any data which would infringe the rights of riders, such as individual rider ratings, feedback, and complaints”.
The ADCU together with Worker Info Exchange brought an additional suit against Uber in October 2020 on behalf of eight drivers from the UK, the Netherlands and Portugal who they claim were algorithmically fired or “robo-fired” for unspecific allegations of “fraud” that they each deny but have had no meaningful opportunity to appeal. Uber has countered that at least two humans review any decision to deactivate an account in Europe.
The three cases the ADCU and Worker Info Exchange have brought in the UK and in the Netherlands could have lasting consequences for the rights of app-drivers and other gig workers beyond Europe’s borders too. As the conditions of workers become as dependent on their data contexts as their physical conditions, the parameters of digital rights movements must shift as well. Or, as Farrar elaborates, “If we can’t access our digital rights, we won’t be able to access our worker rights.”