London-based Uber driver Abdurzak Hadi is self-employed but says he is not his own boss as his workflow is determined by an obscure computer algorithm.
With the ride-hailing app deciding which - and how many - clients he gets each day, Hadi says he cannot optimise work and make more money, prompting him to join legal action against Uber that could set a precedent for all workers in the gig economy.
"I'm blind. I don't know what's going on. I can only see what the app gives to me," said Hadi, 41.
He is one of 10 drivers to sue Uber, demanding full access to their personal data and insight over the workings of the company's algorithms, in a case that lawyers say could improve labour conditions for millions of gig workers in Europe.
Uber is part of a constellation of companies that consider themselves "platforms" - where workers can pick up work in a flexible manner on their own terms - and are booming as rising unemployment pushes more people into the gig economy.
The lawsuit, filed in July and expected to be heard this week, is one of two brought this year against Uber in the Netherlands - where the US company has its international headquarters - while other companies have faced similar complaints from gig workers about data rights.
The lawsuits come as an informal global movement of gig workers has expanded in the coronavirus pandemic, with drivers and delivery workers from the United States to India staging strikes to demand better pay and working conditions.
Matthew Feeney, director of the Project on Emerging Technologies at the DC-based Cato Institute, said people left jobless by economic shutdowns have sought a lifeline in parts of the sector that are booming, such as food delivery services.
"It's good during these difficult economic times that the gig economy offers people a way to make some money," he said.
But the pandemic has also increased and normalised the level of monitoring workers are subjected to worldwide, researchers and advocates say, with a flurry of apps and software measuring how productive they are or what they get paid.
Jeremias Adams-Prassl, a University of Oxford researcher who focuses on the gig economy, said that "what we are seeing now is you can exercise a hitherto unprecedented control over workers".
"Uber can write software in California, and control drivers from Jakarta to Cape Town," he told the Thomson Reuters Foundation in a phone interview.
Those concerns have spurred a growing body of research analysing how gig economy apps behave on workers' devices, allowing people like Hadi to seek more control over their data.
"They (Uber) process data, they can use data to track behaviour, and drivers individually are very much dependent on Uber for their work, they don't have any bargaining power," said Anton Ekker, a Dutch privacy lawyer representing the drivers.
Uber spokesman Rick Janse Kok said that "our privacy teams have worked hard to process all of the data and information requested by these drivers".
The company only uses "factual indicators" to match drivers and riders, he added, such as proximity, direction of travel, and any preference for premium trips, he added in emailed comments.
Uber has pledged to improve the quality and security of platform work, including making its earnings estimator available to more than 40% of active US drivers in over 30 cities by the end of the year.
Before the pandemic, the size of the gig worker economy had been projected to increase from 43 million workers in 2018 to 78 million in 2023, according to a report by debit and credit card company MasterCard.
Although experts say it is difficult to measure the size of the sector during Covid-19, a June report by research agency Gartner found that nearly a third of employers were already replacing full-time workers with gig workers and contractors.
But the rapid growth of the sector has fuelled higher scrutiny from rights groups and complaints from workers over labour conditions.
The Uber workers taking legal action in the Netherlands say the company profiles drivers under a number of metrics, including customer ratings and cancellations, which inform algorithms that allocate jobs.
Accessing this data is key for the drivers to make informed decisions on how to best run their business, said Hadi, such as where and when to pick up customers, or whether it is worth taking out a loan to finance a new vehicle.
But earlier information requests filed by the drivers with Uber under European data regulations yielded little data, say the claimants, who are asking Amsterdam's district court to force the company to hand over any data it has about them and reveal how it is used.
Uber's Janse Kok said, "the information they (drivers) are entitled to was sent promptly but we would never share any data which would infringe the rights of riders, such as individual rider ratings, feedback, and complaints".
Other companies have faced similar complaints from workers about data collection and how it is used.
In October Shipt, a grocery delivery platform owned by US retailer Target, faced a backlash after it changed the way it calculates pay for drivers, shifting from a commission system to one based on an algorithm that the company says measures the "effort" workers put into the job.
"They say they are paying on 'effort,' but we don't know what variables are involved, what they are really measuring," said Willy Solis, a Shipt worker from Dallas, Texas, by phone.
Shipt dismissed the concerns, saying in emailed comments that the new payment system did not change the average pay for workers, and they "always have a choice when they review an order, they get to decide to accept or decline".
Adams-Prassl at Oxford said algorithms are increasingly used not just to set pay rates, but also to make hiring and firing decisions which could in principle lead to fairer decisions.
"We have unlimited empirical evidence that human managers setting pay are terrible; just look at gender pay disparity," he explained. "So you could conceive of an automated system that does a much better or fairer job."
But the trend also raises concerns about transparency and discrimination, he added.
In October Ekker filed a second lawsuit accusing Uber of "robo-firing" - the practice of algorithms single-handedly deciding dismissals - which he says is illegal under GDPR.
The case brings together another four drivers from Britain and Portugal who say Uber sacked them over the past two years, with no right to appeal after its algorithm detected fraudulent activity - including "irregular trips" - which the drivers deny.
The drivers are asking a Dutch court to overrule the firings, arguing they were given little explanation of what they had done wrong and were provided with no supporting evidence.
Yaseen Aslam, head of the British App Drivers & Couriers Union (ACDU), which backs the two drivers' lawsuits, said that "we are seeing a glimpse into an Orwellian world of work where workers have no rights and are managed by machine(s)".
Uber has denied account deactivations are decided solely by a machine, saying "at least two specially trained members of the Uber team review all facts and circumstances in every case before reaching a conclusion to deactivate a driver in Europe or not".
While involving only a small number of Uber's almost 4 million drivers worldwide, Ekker said the two cases could "set a precedent for other drivers but also for other platforms and all people in the platform economy".
Technology tools developed by gig platforms to manage their workers en-masse are migrating to other parts of the economy, said Adams-Prassl.
"We are seeing big data used in human resources, in recruiting, and firing, and retention," he said.
Pathlight, a US technology start-up that allows call centres to manage employees through an algorithm that assigns scores to workers based on performance, has seen a threefold increase in global usage since the pandemic.
It measures indicators such as how quickly workers move through tasks, said CEO Alexander Kvamme. Both workers and bosses can see the score in real time, allowing workers to adjust quickly to meet their manager's expectations.
"It's a weather app for performance," Kvamme said in a phone interview. "Our goal is the empower the frontline worker to take control."
In India, major gig platforms have made it mandatory for delivery drivers to install the national Covid-19 contact tracing app Aarogya Setu.
In June the Indian Federation Of App Based Transport Workers warned that the data collected by the app could be used in a "retaliatory or exploitative manner", including to target workers who might try to unionise, by tracking when they meet with each other.
Urvashi Aneja, founder of Tandem Research, a technology and policy group in Goa, India, said that "now because of the pandemic, you have location tracking on all the time - regardless of if you are working or not".
Companies say this is justified on safety grounds.
Indian food delivery app Zomato said in emailed comments that "the idea is to keep individuals, as well as the authorities, informed in case our delivery partners have crossed paths with someone who has tested positive for coronavirus".
Some groups are trying to help gig workers use monitoring tools to their advantage.
Dan Calacci, a researcher at the Massachusetts Institute of Technology's Media Lab, is building technology tools - including apps and data processing systems - to analyse how gig economy apps behave on workers' devices.
"That way labour organizers, or workers themselves, can see what info the apps are collecting on them - and make better decisions," he said.
Calacci has already coded a free chatbot for Shipt workers, which allows them to automatically analyse their pay stubs under the new algorithmic payment system, in order to calculate the change in pay-rate.
As part of their data request in the Netherlands, Hadi and his fellow claimants are asking that the information from Uber be ported directly to Worker Info Exchange, a data trust set up by their union to build collective bargaining power.
Established in 2019, the trust aims to pool data from workers across different platforms and analyse it to inform more legal actions, said its director, James Farrar.
Insight on how much drivers make per hour spent on the road for example could be fodder for minimum wage claims, he said.
Hadi, the driver, says his salary has plummeted by 30% since 2014, and after costs he now earns on average less than Britain's minimum wage of 8.72 pounds ($11.70) per hour.
Uber's Janse Kok said "drivers choose to work in very different ways and this makes it hard to say what a given driver is earning at any point in time".
A recent study the company participated in with the University of Oxford showed that drivers in London were earning "well above minimum wage, making on average 11 pounds per hour after costs".
Insights on data processing could be used to argue gig workers are employees rather than contractors, said Farrar.
Gig sector companies won a major victory in California in November, after an industry-backed measure known as Prop. 22 successfully passed a state-wide vote, cementing workers' positions as independent contractors.
Yet in Britain, an employment tribunal ruled in 2016 that Uber drivers were entitled to workers' rights such as paid holidays and rest breaks in a case brought by Farrar that is now before the country's Supreme Court.
In September Spain's Supreme Court ruled that riders for Barcelona-based food delivery app Glovo were employees, not freelancers.
But bringing such cases is becoming increasingly difficult with automation, as management decisions are hidden behind algorithms' digital codes, said Farrar.
"Now there is no need for messages, for 'tea and sympathy', for HR and change management - we will just change the algorithm, the processes, and you'll either starve or adapt," he said.
In November, the British Trades Union Congress (TUC) announced it would launch an AI taskforce to push for more transparency and investigate potential tech discrimination in the workplace.
Back in London, Hadi, who has three children and said the pandemic has made clients even scarcer, hopes the court case will hand him tools to better understand how to balance work and family life without losing out on money.
"The reason for all this is that we need transparency."