Uber knows. Do you?

With the example of Uber, I illustrate how law shapes and is shaped by the contemporary, dominant economic system and, thereby, co-constitutes  its core features, such as, work, market and social hierarchies. Responsible legal research acknowledges this. LPE research concretizes it and analyzes if and how law could be used to co-organize a more emancipatory present and future.

Real-Time-ID-Check

One spring evening in 2021, an Uber driver in the UK, Pa Edrissa Manjang, came back home from work. He checked his emails, among them one from Uber declaring: “Due to safety issues regarding account sharing we made the difficult decision to end our partnership with you. We understand that this news can be upsetting but our decision is final”. Pa Edrissa Manjang had been fired based on the allegation that he had shared his Uber account with someone else and there was nothing he could do about it.

Pa Edrissa was indeed upset. When he got fired, he had already been working over a year for Uber and had never shared his account with anyone. Instead, what Pa Edrissa actually had experienced, was a lot of trouble with a facial recognition system integrated into the UKs Uber “Driver App” in 2020 called “Real-Time ID Check”. This system, however, did not recognize him on several occasions – which is why he regularly had trouble to verify his account.

With the Real-Time ID Check Uber verifies if a person is a registered Uber driver. This is supposed to guarantee safety for Uber passengers. The Real-Time ID Check works as follows (Microsoft): when an Uber driver registers into the app, they upload a photo of themselves. The facial recognition system creates something comparable to a machine-readable copy of this photo. Any time the Uber Driver App asks an Uber driver to verify their identity (which happens randomly), they take a selfie and upload it to the app. The facial recognition system again creates a machine-readable copy and compares this new copy to the registered one. If these copies match you are a registered Uber driver; if they do not match, you are not.

Face too Dark

The facial recognition system integrated into Uber’s Driver App is provided by Microsoft Azure. Microsoft Azure’s facial recognition system was part of Joy Buolamwini’s and Timnit Gebru’s “Gender Shades” study published in 2018. The two computer scientists audited several commercial facial recognition systems and found out that every system worked worse for Black men and especially bad for Black women; Black men and women were misrecognized at a higher rate in comparison to white men and women. The US National Institute for Standardization and Technology came to a similar result in 2019. While Microsoft Azure commits to make their facial recognition system more accurate for everyone, their endeavors apparently continue to fail: in 2021 the computer scientists Gaurav Jain and Smriti Parsheera published a study showing that Microsoft Azure’s facial recognition system was still worse at recognizing Black and brown people.

In light of this information, it comes as no surprise that Pa Edrissa, who identifies as Black, was misrecognized on several occasions by Uber’s Real-Time-ID-Check. While his white colleagues had to verify their account once a week, he was asked to upload a selfie up to three times a day (Manjang). Other Black or brown Uber drivers in the UK, the US or India were likewise fired or their access to their Uber account was temporarily blocked for the same allegation of an account sharing that had never happened. Each of them lost their jobs – at least temporarily – because of a facial recognition system that did not and still does not work for them. Pa Edrissas story is not an exception to an otherwise working system.

Work

The Real-Time-ID-Check is part of Uber’s broader control and surveillance infrastructure of their drivers. Whereas Uber subjects its drivers to a system of hypervisibility and transparency, Ubers decision-making structures and its effects remain opaque. Uber drivers are not told how Uber’s platform works; why they are sent to one part of the city instead of another, how they are rated, why they must verify their account three times a day, why exactly they were fired (Manjang; Bansal).

Some Uber drivers around the world share their experiences online and try to reverse-engineer the platform’s reasoning (Bansal). Within the UK, this collectivization of Uber drivers’ experiences is facilitated by a wide field of unions, such as the Independent Workers Union Great Britain or the App Drivers and Couriers Union. But sharing experiences and more transparency alone, do not automatically lead to participation or change. To participate, that is to be able to contest and co-decide about working conditions, requires participatory venues. Uber offers an email address.

Compared to an email address, rights, such as the right to non-discrimination, offer a more powerful venue for participation (although mainly focused on mitigating harm). Uber knows this. It was only in 2021 (after more than 5 years, that Uber already operated in the UK) that the UK Supreme Court ultimately decided that Uber must guarantee some workers’ rights. A claim that Uber had tried to rebut over several years.

While this legal struggle was a success for Uber drivers in the UK, to actually enforce these rights, especially if technology is included, faces many barriers. The requirement of technological knowledge, time, and money being some of them. Pa Edrissa Manjang was able to overcome some of these barriers. Backed up and financially supported by the App Drivers and Couriers Unions as well as the Equality Human Rights Commission and based on shared experiences as well as on Buolamwinis and Gebrus research, Pa Edrissa filed a complaint of discrimination with the London Central Employment Tribunal in 2021. After three years, his complaint was settled out-of-court. He got a payment. No Court ever decided whether the facial recognition system had discriminated against him or not; if Uber needed to change anything in this regard or not. As BBC reports this “out-of-court settlement marked the end of a ‘long and difficult’ period” for Pa Edrissa. To enforce rights also requires a lot of resilience – especially in face of a powerful player such as Uber.

Market

Pa Edrissa, now, works for UberEats. He did not share his reasoning publicly. I can only assume that it may also be due to a lack of options. Within the UK, UberEats holds quite a share of the food delivery market. In Germany, UberEats is not as popular but among other delivery services, they are a quite established part of German streets scenery and people’s daily live. This is worth mentioning because it is not the same for Uber’s passenger transport service.

While Uber is a big player within the UK’s passenger transport market, within Germany they are not. Even within the EU, Uber’s share on the domestic passenger transport market differs. These different market shares are not the mere result of different national demands, but also of different laws. In 2017 the European Court of Justice decided, to the detriment of Uber, that the EU did not harmonize the market for Uber’s transport service. Therefore, Uber must comply with divergent Member states’ passenger transport laws. Political economists have shown how these different laws shaped Uber’s passenger transport business and market access differently (see e.g., Thelen comparing Germany, Sweden and the US; Pernicka/Schüßler comparing Germany and Austria; for a non-comparative but closer analysis of the US market see Dubal).

In Germany, for instance, Uber’s service UberPop was prohibited by a Court ruling in 2015. Nonetheless, Uber’s passenger transport service is still offered in a few German cities. There, Uber’s passenger transport service is no longer running as a taxi but a car rental service which entails new legal and city-specific, administrative struggles over a so-called obligation-to-return (Rückkehrpflicht; Federal Court on UberBlack in 2018). This legal construct, on the other hand, shapes Uber’s passenger transport service in Germany differently to, for instance, their service in Austria, where Uber needs a license as travel agency. In June of this year, Uber announced that it is expanding its passenger transport service Germany-wide – with a new legal arrangement.

Social Hierarchies

In London, Pa Edrissa is one of many Black persons, who drive for Uber’s services. BAME (Black, Asian and Minority Ethnic) form the majority of Uber drivers in London. Also in Germany, it is especially non-EU-migrant persons who work for platform services (Schaupp. A trend in other countries, too: van Doorn/Ferrari/Graham). It is them, who make platform services work, in the first place.

In Germany, this is not a coincidence. The political economist Simon Schaupp shows how migration laws in Germany co-constitute a migrant low-wage-workforce by strongly restricting the access to the regular job market for non-EU-migrants and leaving only a few openings for precarious, low-wage-work for them. He also shows, how platforms and other algorithmically-managed work now allow employers to profit from this pool of legally co-constituted low-wage non-EU-migrant workers as the language of the corresponding platform can be adapted and no contact is needed between employers and workers. While these laws and technologies create jobs for people who cannot have one (which is, again, also an effect of restrictive migration laws), law and technology are, at the same time, used to segregate the work market even further.

Some of these workers experience solidarity among each other and would like to improve their working conditions together. However, in Germany, this remains difficult. Given that in Germany laws restrict the right to unionize and strike (Arps/Tügel; Haufe; Kocher) and given that migration and low-wage-employment are both fields with little possibility for political participation, rights often remain the only avenue for legal contestation for people otherwise assigned to the bottom of society. But, even if supported by many, rights are very difficult to be actually enforced as the example of Pa Edrissa Manjang has shown.

Responsibility and passionate reconstruction

The case of Uber illustrates that law goes way beyond mere conflict settlement in an individual case. Law shapes and is shaped by the contemporary, dominant economy and, thereby, co-constitutes its core features such as work, markets, or social hierarchies. Worker/non-worker, harmonized market/non-harmonized market, taxi service/car rental service, migrant/citizen etc. are not natural phenomena. Law, in its own legal manner, co-constitutes them and the social consequences that follow (Maihofer 1993; Latour 2005; Buckel in Buckel et al 2022).

Uber knows this. They rely on their technology, cultural habits (getting used to their technology), and legal struggles to establish the workers they need (without workers’ rights) and the markets they wish. Through commentaries, articles, Court proceedings etc., we – legal scholars and lawyers – are part of the process. There is no idealistic, objective “nowhere” from which legal scholars could produce knowledge. Responsible legal research acknowledges this.  To be responsible means to reflect the ways we learn to see, to not naturalize but explain, to take a stance, and, thereby, remain responsive (Haraway 1988).

There are, indeed, different ways where this responsible legal research could take you, depending on your theories and methods. To me, LPE offers materialistic theories and methods that analyze how the economies’ core features are institutionalized within law and that analyze, if and how law could be used in a particular setting (e.g. platform work, housing etc.) to support more emancipatory presents and futures (cf. Kampourakis 2021). LPE offers a knowledge production, that contests, deconstructs and passionately reconstructs “worlds less organized by axes of domination” (Haraway 1988: 585). In the field of platform work, there is much to be done for lawyers and legal scholars (for Germany see e.g., the work of Kocher and Hensel or here).