The Gig Economy Project speaks to Jessica Pidoux, director of PersonalData.IO, about her critique of data in the platform economy and how gig workers can use data tools to build workers’ power (in podcast and text form)
THEY call data the oil of the 21st century. Without it, the digital economy as we know it would grind to a halt. And just like oil, unrefined data is pretty useless – it’s what you do with the data that makes it valuable.
What is done with data, and who has power over it, is what preoccupies Jessica Pidoux, director of the NGO PersonalData.IO and a lecturer in the sociology of data at SciencesPo university in Paris.
Pidoux has studied the algorithms of match-dating apps like Tinder, finding they re-produce sexist and patriarchal practices. And she has also applied her data techniques to the gig economy, working, along with her colleagues at Hestia Labs, with an Uber driver in Geneva, Switzerland, to analyse his data at work and understand what information the platforms are keeping to themselves, as we wrote about on the Gig Economy Project (GEP) earlier this week.
Speaking at an event in the European Parliament in September on alternatives to Uberisation, Pidoux said: “Workers can take back this data power.”
In this podcast, GEP speaks to Pidoux to find out more about her critique of data power in the platform economy, what she thinks the alternative is and how she is helping workers build their data power. We discuss:
01:02: Why does data matter and what’s the problem with the status quo?
03:33: The case of Cabify and the Catalan Minister of Transport
07:03: What’s the alternative to a platform-dominated data economy?
09:24: An Uber driver in Geneva’s pursuit of his data
19:20: The Rider’s Law and trade unions accessing the data
22:30: Pidoux’s academic work and Tinder investigation
28:35: How PersonalData.Io can help you access your data
You can read a text version of this interview below:
GEP: Pretend for a second I’m a Deliveroo courier or a Care.com home carer. Why should I really care about my data, why does it matter, won’t my life be easier and more convenient if I just ignore it and let the platforms do what they want with it?
Jessica Pidoux: The problem is that now we don’t have visibility on how the platforms and algorithms work, and the value of the data is that we can gain transparency over the system. And it’s also a means of accounting; it’s a way for gig workers to track their work activity and verifying whether a platform is paying correctly what they are owed.
GEP: At PersonalData.io you argue that the data which is collected on us as workers or consumers by platforms is our data and that we have a right to it, but haven’t we all signed an implicit or sometimes explicit contract to hand over our data to Google or Facebook or Uber in return for being able to use their services or work through their app?
JP: It is a problem that our behaviour is tracked and turned into data, and this data is used as a raw material, a means of production for the platform. So the platform is making money out of our behaviour and data. It’s the only way the company can provide a service, and it’s the only way the company can make revenues because they are also trading this data. The issue is that the platform doesn’t recognise their dependency on the worker and his/her data, and they are not getting back what they should for providing all of this data.
GEP: The Gig Economy Project reported on a story in September in Catalonia, which you commented on at the time, and I think it was quite eye-opening about the way companies can seek to use their control of data in ways that are potentially dangerous to democracy.
Cabify, the Spanish private hire platform, sought to undermine a new law regulating private hire platforms in Catalonia by running an advertisement campaign with figures claiming to be the number of trips made by the Minister of Transport to the Transport Ministry using Cabify, with the idea being that the Minister’s law was hypocritical. Now, the actual data Cabify used referred to all trips to the Transport Ministry by anyone in 2022, so it could have been anyone, not necessarily the Minister of Transport at all, who almost certainly uses a government car to get about not Cabify.
But the point is that here was a company purporting to use the personal data they held on a Minister in order to attack a law made by that Minister which directly affects the business of the company. Is that an example of how data can be used by platforms in ways that are dangerous for democracy?
JP:Yes because data concerns everybody. These platforms are collecting data about the drivers, the riders and the clients, which can be policy makers, parliamentarians, and in this way the platform can understand the offer and the demand. They are tracking the behaviour of the clients and the drivers to understand the market and influence the prices in the market. This means that the issues around personal data concerns everybody, not only gig workers. Politicians are also affected by personal data and the way it is used.
What I think is relevant with the Cabify advertisement is that it shows all the problems we are having today with these data collections. It’s two-sided, the platforms can influence politicians that are involved in policy-making, who are actually also controlled because their attention can be influenced and their behaviour can be influenced. And on the other side the platform obviously influences the gig worker, because they can choose who gets the best rides, who gets the best prices, the best clients, or not, who gets banned and often without giving any explanation.
GEP: What are the alternatives to an economy based on data being owned and controlled by mega-corporations?
JP: This question allows me to complete the previous one. It is a problem for democracy the way data is collected and used, and at the European level and at the national level these platforms are not controlled and the regulations can be hardly implemented because the policy-makers and the state lacks visibility on how platforms work.
And then what I see as a new alternative, new possibilities of innovation is that if the state is today giving a lot of liberty to companies to innovate and contribute to the economy, this actually should be given to society, to social actors, to gig workers, to redefine how this data is used for their own good and for the common good, even for the common good of clients that can be affected by the prices which the platforms arbitrarily assign.
So for me there is relevance for developing collective governance, helping the gig workers to have technological infrastructures. They need the technological resources that now only dominant platforms have. And when we give them these resources and a space to discuss and make decisions together, then we can plan new ways of making innovation and taking decisions for the common good.
GEP: Ok, let’s talk about some of the work you are involved in with respect to data and data power. We’ve just published a piece on the Gig Economy Project about PersonalData.IO’s collaboration with an Uber driver in Geneva, Switzerland, Guillaume, where you have helped him try to access his data and written up a report about what you were and weren’t able to find out.
One of the things that I think was interesting about that is just how difficult it is for a worker to access their data, even though they legally have the right to under EU GDPR laws, in a meaningful way. Can you just talk us through the process that you went through?
JP: We all have data access rights. Thanks to the GDPR and data protection law any person can request a copy of their data to a company. So it can be a supermarket that digitises your consumer behaviour, it can also be to a platform like Uber.
So we used those rights to make a Subject Access Request (SAR) and a data portability download with the Uber driver to get his data. He did it in 2017 and he made another request in 2022, where actually we found out that Uber doesn’t give all the data. And the process is that we use a letter that we provide to Uber drivers and any gig workers based on the law, so you request your data, and when you send this letter the platform has a maximum of 30 days to respond and give you your data.
What happens is that Uber start finding work arounds to not give you the data, and the problem is that the quality of the data they give cannot be analysed afterwards or that the data we get is not complete. One example is geolocation that is time-stamped, which means you get a specific geographic point where the driver is connected, and we can see the date and the time when this happened, and Uber only gave the last two months of data. Which means today, given the problems that are facing the drivers in Geneva where they are now recognised as employees and they want to calculate the revenues that Uber owed them, but without this geolocation data it is not possible for the driver to calculate the revenues that Uber owed them.
Uber says ‘well it’s a problem of privacy, we need to protect others personal privacy, and if we give you all this geolocation data we might put in danger other people, because you have with geolocation the home addresses of the clients, where you left the clients’ and so on. But actually, the law says that Uber or any company should provide the means to give you this data in a protected way. You can anonymise the data, that means that you don’t give the name that is associated to the data, or justify why it is a problem, you can divide data types, so that you don’t combine the driver’s data with the passenger’s data, to protect the knowledge that you can extract from this data. And Uber said ‘no, we cannot give you this, but give us a specific period of time and we will give this data to you’. And we asked now for a specific period of geolocation data and we are awaiting the answer.
This is not the first answer like this that we have got from Uber and other platforms. They make you request permanently to avoid giving you a copy of the data. And I think it’s important to secure data and protect everybody’s privacy – it is a reality that you cannot give all the data that is possible to collect in a platform. But if you have a good internal process and control of data, then you should be able to provide this. They have good technologies now for predicting rides, they have sophisticated resources so they should also be capable of providing drivers and gig workers with their data. What we have seen is the files they return are a complete mess, sometimes they give geolocation data in miles instead of kilometres, there are missing rides for the driver so he doesn’t have data on certain days that he worked that he should have, etc.
GEP: One of the interesting things about this is that it feeds into an ongoing dispute between Uber and drivers in Geneva over back-dated pay, after a Court found that these drivers were employees and thus Uber owed them years worth of salaries. But what exactly they are owed has been a source of disagreement, but how can you come to an objective perspective on what they are owed without being able to see all the data, right? So PersonalData.Io has been trying to convince the workers’ lawyers to take the data issue seriously, is that right?
JP: Yes exactly, we are saying that data can be proof of working time, and this can help you make the calculations and find a method to negotiate with the state and with Uber in the court case to reclaim more money.
But the problem is that now the gig workers are supported by trade unions where there is no expertise on data, they were not aware or didn’t care about recovering the workers’ data until very recently. The state does not have experts on data, algorithms and how platforms work either.
Everything in the current negotiation is around work law, and what we are saying is you need to include the data protection law and the issues around transparency of the data that is collected because this is the only way to track how the work is done. Some unions are asking the drivers ’oh did you have an accounting of your hours, did you write it on a paper, or an excel file, or did you use a disk in a car that tracks all the kilometres that you did?’. These are the only means that are proposed to gig workers and what we are saying is your data is the proof of your work and you have to recover it.
Also the reality today is that Uber is not meeting the expectations of [GDPR] law, and this can be used also against Uber, because they are not providing the proofs that Uber drivers require. You have a legislation that can support this, so what we are saying is these cases, all around the world, need experts on data and need data protection lawyers.
GEP: In Spain, the ‘Rider Law’ was passed last August, which did two things: it gave a presumption of employment to food delivery couriers, but it was also supposed to allow worker representatives, including trade unions, to access the algorithm of these platforms, but since the law was passed over a year ago we haven’t heard anything about those powers being used until this week, when one of the big unions, CCOO, they’ve released a press statement saying they are making an application now to access this data. That’s about 14 months after the law was passed, so I guess that reinforces your point that it is one thing to pass laws for data transparency, but if you don’t actually have the people in place who can do it, then you can’t really act on those legislative changes to make them a reality for workers.
JP: Yes exactly because everything we are claiming is algorithmic transparency and if we don’t know how to request that and don’t know how to audit that…[then it’s a problem]. What we are saying is that is that you need to get a copy of your data because this will empower you, it will give you knowledge of your activity. And it’s a pretty easy and implementable way to gain autonomy over the platform, but for that we need the support of the state and the regulators to actually ensure that these data access rights are really actionable.
What is happening today is because people are not aware of these rights they don’t request their data, some workers are afraid of requesting them, because sometimes they think ‘well I worked extra hours that were not legal, I will be punished by the law’. But actually if you go to a court case they will not give this information to the police and denounce you, so actually there is a quick, easy way to gain power over the platforms but it has to be supported by initiatives from the state.
Because also what we have seen from the Uber driver case in Switzerland is that the company is not compliant with the data access rights and when we complain to the competent authorities for data protection they say ‘well we have a lot of work we cannot treat these individual cases, talk to you later’, basically. So we are not getting the support of the authorities to recover this data, so if we put more effort and initiative here then we have an immediate short-term solution to gain transparency and make these calculations.
GEP: The other thing you do is you’re a sociologist and you lecture at SciencesPo in Paris, trying to engage students in issues around data and society. Tell us a bit about your academic work?
JP: I’m a lecturer at SciencesPo in Paris in the European study programme with a focus on digitalisation. It’s a Masters programme and it’s actually a super nice experience for me because these students are the future policy-makers in Europe. So what I do there is I teach sociological concepts and theories but also practical case studies, for instance of the Uber driver in Geneva but also other cases, and together we analyse these cases from theories of sociology to understand that technologies work in a very precise way and in a very technical way so these policy-makers can be better informed in the future and can make better decisions on how to regulate platforms.
That’s what I do, I combine sociological concepts and data techniques, and the students analyse their data too, they recover it from different platforms – Twitter, Uber – and we situate their knowledge on the data they have recovered, on the process of recovering the data, and we consider all the perspectives that are important in these cases, which is the user of Uber but also the driver, etc.
GEP: You are the first researcher to have studied the Tinder dating app algorithm, and your research on this was the basis for a book called ‘Love under algorithm’. Tell us a bit about that?
JP: I did my doctoral thesis in digital humanities and my dissertation was focused on algorithm dating apps, which actually brings me now to be very interested in Uber and these kind of platforms because there is also a matching process, you need to profile Uber and the clients, and you need to match them, with the idea of optimising the best distance between them and the best prices for the company to make profits and potentially for better revenues for the drivers.
So I analysed more than 30 dating apps, and one of the biggest in the market is Tinder, and I analysed the patents of this app, and this is also something I do today with PersonalData.io, we make audits of algorithms and data collection processes.
From this patent of Tinder, I showed that all of the data that is collected – your swipes, your likes, your age, your geolocation, your sexual preferences, your education level – is used to present your profiles, and I show that Tinder uses this to favour a patriarchal model. This means that men with a high education level and a high income, they are actually introduced to women that are younger than them and have a lower education level. So I show how the design of these apps and algorithms introduce biases from the beginning, because they decide that some information is more important than others, and actually an asymmetric relationship is favoured between women and men in this case, because then you can re-produce social norms that are in society.
In the case of Tinder, a lot of heterosexual people are registered and actually how they use the app might be in accordance with this; women might prefer older men. The problem is the app reinforces these stereotypes in society because they re-produce this model of the majority with anyone and then they discriminate against minorities, they don’t consider diversity, they don’t consider equality between sexes.
GEP: Fascinating, and if people want to find out more what should they do?
JP: I would recommend to go to PersonalData.Io, you can find ways to recover your data methods, tutorials, and we also have a youtube channel, and then we have a wiki where I explain more about dating apps, and it’s called wiki.datingprivacy.
GEP: If a gig worker is listening to this and thinking, ‘you know what I want to access my data’, can PersonalData.Io help with that?
JP: Yes, totally, our mission is to help gig workers and civil society to become experts on data so they gain new skills to control their activity on the platforms – their work, their lives. And I just ask them to write me a message, jessica@PersonalData.IO. We are also doing workshops and all of this can be found on our webpage.
To sign up to the Gig Economy Project’s weekly newsletter, which provides up-to-date analysis and reports on everything that’s happening in the gig economy in Europe, send your e-mail to GEP@BraveNewEurope.com.