Faculty of law blogs / UNIVERSITY OF OXFORD

A Book Review: Humans as a Service (OUP 2018) by Jeremias Prassl

Humans as a Service: The Promise and Perils of Work in the Gig Economy (OUP 2018) by Jeremias Prassl is an engaging, illustrative, and thought-provoking book. The book argues that gig-economy platforms such as Uber, Amazon’s Mechanical Turk (MTurk), Deliveroo, or TaskRabbit ‘present themselves as marketplaces, even though, in reality they often act like traditional employers’ (ibid 5). Unveiling this doublespeak of the gig economy in which the platforms ‘operat[e] under the mantle of a “sharing economy” [but in fact focus] … on commercial labour intermediation’ (ibid 7) is one of the main targets of the book, because, as Prassl argues, the doublespeak is ‘designed to keep customers and workers interested, and regulators at bay’ (ibid 85).

According to the book, one of the issues brought about by the gig economy is a seemingly new class of workers and employers who do not easily fit the traditional concepts known from employment law. The book asks whether the gig economy micro-entrepreneurs, ie the Uber drivers, Deliveroo couriers, or Amazon MTurk workers, should be characterised as employees and the online intermediary platforms as their employers. This issue is not new and has been already discussed in courts, by the lawmakers, in the news, as well as in the academic literature, as we can see from the wide range of international sources on which Prassl’s book relies. Instead, the main original contribution of Humans as a Service rests in the following three claims that the book advances in response to the question.

Three original claims

The first and, according to the author, ‘[t]he central claim of this book is that most gig-economy work falls within the scope of employment law’ (ibid 140). This claim is supported throughout the book by a number of practical examples regarding the functioning of the gig economy, most often by secondary empirical data regarding operation of Uber in the UK and US. This intriguing collection of real-world cases gives a vivid picture of the underlying processes in the gig economy. Besides, Prassl’s book guides the reader through an engaging and convincing historical journey to show that ‘platforms have latched onto our deserved support for innovation—and sold us a century old model as a radically new way of organising work’ (ibid 85). The most fascinating data concern the lock-in mechanisms which the platforms use to gain and retain a substantial (or dominant) market share, hence leaving little freedom to the gig workers as to whether they will sign out from the app, or just turn it off for a moment.

Despite the proclaimed ambition to describe ‘the reality of work in the on-demand economy’ (ibid 50), however, a number of important empirical claims in the book would gain in persuasiveness with some more backing from empirical evidence. For example, the book does not substantiate the claims that workers ‘suffer precisely because supply is designed far to outstrip demand’ (ibid 21), or that ‘cars [are] roaming the streets whilst looking for the next passenger’, which has environmental impacts (ibid 21). Other important claims seem a bit overstated. For instance, while the book shows that some workers are ‘tightly controlled by their platforms, with low pay, long hours, and questionable working standards’ (ibid 68), it is not clear how the same evidence proves that these ‘working standards … are increasingly the norm, rather than an exception’ (ibid 68). Besides, the book’s overall argument would have been stronger had the occasional overgeneralised interpretations of data also been avoided. For example, a statement that ‘only a little over 50 per cent of registered [Uber] drivers are still active a year after they first signed up’ (ibid 68) only applies to US Uber drivers who made their first trip between January and June 2013, ie more than five years ago, as one can see in the original source of these data. Still, the book’s claim that ‘the reality of work in the on-demand economy can be so disheartening that workers move on to different platforms’ (ibid 68) relies on this interpretation. As a result, it is sometimes unclear about what regions or markets, for which period, and about what domains of the gig economy the individual claims and hence also the book’s conclusions are made.

The second and pehaps also the most important original contribution is that the book convincingly argues against the ‘all or nothing’ approach to the employee status of workers in the gig economy (ibid 101). The dilemma, as it is often seen, is that the gig workers are either employees providing a controlled service or independent contractors (ibid 44). This apparently leads to conclusions that the gig workers who do not easily fit the concept of an employee must be seen as independent entrepreneurs, or vice versa.

The book nicely shows, however, that the reality is more complicated and that we need a case-by-case approach when determining the status of each worker in the gig economy. Importantly, in this regard, the book argues that if we focus on the entire customer-worker-platform relationship we can often find all the main functional aspects of the traditional employment. Thus, although it might be hard to present the gig workers as employees in the traditional structure of the two-party employment contract, the book makes a strong case for adapting employment law to this new structure and for granting gig workers employment rights and legal protection. An important issue that the book leaves us with some uncertainty about is against whom these rights could be granted. On the one hand, the book suggests that the gig-work platforms should bear the employer’s responsibilities, but, on the other hand, it acknowledges that these platforms do not always concentrate all the functions necessary to characterise them as employers.

Thirdly, after showing that there is nothing new about the business model behind the gig economy—the model still relies on what we traditionally characterise as labour—and after showing that employment law (with further advancements) could apply to the gig economy, the book puts forth a prescriptive claim that employment law should regulate gig-work, because it will help level the playing field ‘to the potential benefit of all involved’ (ibid 140). According to the book, employment law should regulate work in the gig economy because ‘[c]onsumers, taxpayers, and even platforms themselves will ultimately benefit from a marketplace in which everybody plays by the rules’ (ibid 117). Besides, employment law ‘appears to be much the most efficient way of ensuring that employers internalize the costs of their business activities’ (ibid 131).

Although it is clear that employment law is well suited to allocate fairly the true costs of such intermediated labour, the book does not explore any variant regulatory instruments to support the claim about employment law’s supreme efficiency. The book promises to show us benefits that a levelled playing field will bring, but then discusses the costs and negative effects that it could help avoid, which is a different issue. Moreover, the book acknowledges the importance of other regulatory areas of law—especially consumer law and tax law (ibid 131), but not, surprisingly, competition law—that could help solve those issues regardless of whether we characterise the gig-work as employment or not. As a result, it remains unclear why we should buy into employment regulation at large, rather than addressing the individual issues brought about by the gig economy.

Why humans as a service?

The title of the book is inspired by the Amazon CEO’s talk from 2006, in which Jeff Bezos discussed the introduction of MTurk and its impact on cloud computing: ‘You’ve heard of software as a service—Well, this is basically people as a service’, he said (at 11:40). The powerful idea behind the MTurk was that in addition to a cloud of software services that the users may utilise, you could tap into a cloud of people who would provide services to supplement tasks which the computer software cannot perform as efficiently as humans. Back in 2006, for example, these human-reliant services included simple and quick yet intelligence-heavy tasks such as visual recognition. Today, these human intelligence tasks range from writing a new line of code to filling in a survey, from reviewing a new app to editing a translation. The model behind remains the same though—as a consumer you still use the online app which intermediates the demanded service, be it a service performed by a computer software or supplemented by humans.

If we take the book’s title seriously and not as mere doublespeak, perhaps we should think similarly about the gig economy at large. For example, we should think about Uber price-making algorithms as one layer and Uber drivers as another layer in the cloud of Uber’s services. This analogy, however, does not invite so clearly questions regarding the employment status of the workers in the cloud. If we pursue the analogy, it is evident that the workers could be qualified both as employed by the cloud service provider (a platform) and as hired by the end user as independent contractors. Accordingly, the ‘all or nothing’ approach does not play any significant role in this cloud model of gig-work.

For some reason, however, the book pays little attention to this line of thought and does not elaborate on the analogy in its title. That is surprising because Human as a Service (HuaaS) is a concept long employed in cloud computing on which the new platforms largely rely. To the extent that the book wanted to discuss regulation of these platforms, its arguments could have explored the analogy and perhaps point out the dangerous commodification of humans in the cloud. In the absence of such discussion, it seems unlikely that the book will bridge the divide between the language of technology and the reality of the business models that underline it.

Final remarks

Humans as a Service is a great starting point to anyone interested in employment law issues surrounding the gig economy. It provides an authoritative overview of some of the most topical problems and illustrates them by a number of exciting examples. Although the book does not address any specific jurisdiction or segment of the gig economy, it usefully instructs the reader on how to make better sense of the business models, algorithmic governance, and network effects underpinning much of this new domain of economic activity.

Of course, a book of this format cannot provide an exhaustive description of the reality of work in the gig economy at large. Hence, to see whether ‘the real prize for corporate doublespeak has been in selling platform-based work as innovative entrepreneurship’ (ibid 40) and, therefore, whether employment law should regulate this type of work, our labour judges and administrators will need thoroughly to examine individual cases, as the book convincingly argues elsewhere (ibid 100).

Overall, there is a lot of food for thought in this book and its unusually engaging style makes it even more tasteful. Humans as a Service is an important contribution to debates about humans’ labour in the digital economy and invokes a more structured thinking about these issues.

Václav Janeček is reading for a DPhil in Law at the University of Oxford.


With the support of