#8_Digital labour inspection: Just a fancy or a real innovation?

Across the social sciences, the digital economy gets widespread attention for understanding the social organization of work in the present, and even more so in the future. Most of the debate is driven by projections based on the potentialities of the newest generation of information and communication technologies such as artificial intelligence and machine learning in advanced robotics (e.g. Brynjolfsson & McAfee, 2014; Plesner & Husted, 2020; Upchurch 2018). In amalgamation with digitally mediated social networks, cloud computing, and blockchains, and fuelled by the outstanding business performance of US- and China-based tech giants, these technologies are expected to fundamentally alter how we work, when and where (UNCTAD, 2019).

Although a potential for change is hardly denied by anyone, the literature diverges on the extent and forms of change to be encountered. In parts of the literature, electronically mediated work such as gig work, crowd work or cloud work are seen – for better or worse – as a likely future of work. For other observers, the social risks and economic limits of these forms of work foreclose their widespread deployment in all workplaces across the economy. The controversy continues with figuring out the size of the phenomenon: Depending on various estimates between a fifth and a third of all jobs may go “gig” (e.g. Woodcock & Graham, 2020), but more often a share around 10 percent is reported; still a number in the hundreds of millions of people concerned (e.g. Crouch, 2020). One difficulty in estimating the size of the phenomenon consists of identifying clearly the distinguishing characteristics of digital working. For example, separating lines are drawn by the varying degree of jobs’ spatial and temporal boundedness (Woodcock & Graham, 2020), or by the work performers’ autonomy and dependence in carrying out the work tasks (e.g. Kuhn & Maleki, 2017). Such criteria are rough indications at best, largely because also regular jobs are influenced by digitalization (Plesner & Husted, 2020). In other words, not just gig and crowd workers, but also regular employees experience a change in their working conditions through digital technologies. For example, remote work with a regular employment contract is as digitally mediated as is the work of a parcel delivery rider receiving his orders from a smartphone app on the spot. Of course, even where people still commute to a worksite – the Covid pandemic set aside – they often use digital devices for carrying out their jobs from desktop computers down to specialized handhelds. Also, the labour market consequences are controversial: for some a new wave of automation looms around the corner bringing large-scale technological unemployment, for others the effects will be more than offset by the “market creation” capacity of the new technologies (e.g. Autor, 2015, Borjas & Freeman, 2019).

And so far, the qualitative effects on organizing work and job performance remain opaque too. Although it is debatable whether digitally mediated work is always exploitative, for example when workers from countries with large informal sectors can earn a living from online tasks (Wood et al., 2019), oftentimes gig workers are used as an exemplar to highlight the precariousness of digital work (Vallas & Schor, 2020). Obviously, the electronically mediated sweatshop is already a reality for a number of workers (Head, 2014). Labour standards such as working time limits and minimum wages are said to be poorly complied with in platform mediated job areas such as warehouses, logistics, parcel or food delivery as well click jobs; but also in areas where reverse auctions lead to a race to the bottom in prices for professional services and craft trades. Additionally, there are many concerns about labour standards in areas ranging from health and safety, and social security to skill formation, inclusiveness and the protection from electronic surveillance (e.g. for crowdworkers watch Valerio de Stefano).

Unsurprisingly, the issue of labour standards and the huge gaps in their enforcement in digital work has been brought up recently with full force (Choudary, 2018; De Stefano & Aloisi, 2018; De Stefano, 2020). One explanation for the divergence in labour standards between the most extreme forms of digitally mediated work and what is thought of being regular employment is the legal employment status. Inasmuch as a formalization of workers into the status of being employed as a regular employee is key for being protected by labour law, the issue of legal classification of job holders becomes of utmost importance (Cappelli & Keller, 2013; De Stefano, 2020): Are gig workers, or better platform mediated work performers, “employees” in an “employment relationship”? And to what extent are these workers truly “self-employed” persons or just employees in bogus self-employment? Or is platform labour a genuine type of work in need of a new regulation for extending protection? For the individuals working for these platforms and their clients, the question is not so much whether platforms’ business models are championing new technologies, but whether their success is built on a trick which allows circumventing labour standards codified in labour law and collective agreements. This is a complex question because electronically mediated work relationships also include the customers (e.g. Healey et al. 2018). And many platforms appear to use their network power for influencing consumption behaviour and patterns, following a cost-saving approach based on poor labour conditions behind the line of electronic visibility (e.g. Choudary, 2018). For this reason it remains doubtful whether consumers benefiting from the amenities of platform mediated orders may voluntarily become an influential advocacy group for labour standards.

But how to uphold labour standards in this segment of the labour market, then, not least for avoiding negative spill-over effects for the rest of employment, whatever its size? One hope is to use information technology to the benefit of workers and not as a device for worsening working conditions. Such an approach, however, requires a change in perspective towards whether and how the new technologies can be put to a beneficial use for labour standard enforcement. For various reasons, recent literature appears to be hopeful and cautious at the same time as far as the potential benefits of such a change in perspective are concerned. To illustrate why, two examples are briefly discussed from the literature: (1) Changing the governance rules of platforms including setting up worker run advocacy and self-organizing platforms, and (2) using techniques such as machine learning and blockchains for increasing supply chain transparency and network responsibility.

Changing the governance rules may provide real options to protect workers in their position towards the platforms. For example, labour standards could be included in the transparency regulation for platforms in the General Data Protection Regulation (GDPR) in order to protect workers and independent contractors as users of these platforms regarding issues of payment for completed, but refused work results, their rights to access and use their accounts as well getting influence on rating systems (Silberman & Johnston, 2020). The pitfall of this approach is that it is a protection for those self-employed persons to restore freedom of contract in the digital space, but cannot do justice to the power asymmetries in what is a labour contract in effect. Another approach in changing the governance of platform is by building up pressure through advocacy platforms. For one, platform and gig workers can use websites and social media to expose unfair platforms and their practices as well as highlight those who are operating with good labour standards. Such internet platforms may organize like social media platforms and produce ratings and rankings from a workers’ view instead of those given by customers and platforms (e.g. Turkopticon). As such self-organizing is basically driven by campaigning efforts, these efforts are highly volatile in changing platform governance structures from the outside. In cases where the targeted platform is well established and endowed with widespread consumer support for its business model (Culpepper & Thelen, 2020), outside pressure is unlikely to succeed without support from institutions of internal voice. Another variety would be the idea that workers (and end consumers) build coalitions and use platforms to organize the sharing economy business models of the digital age. Compared to advocacy platforms, workers could use their own independent and cooperative platforms for distributing work, thereby upholding labour standards (Charles, Ferreras & Lamine, 2020). Without further organizing efforts such as union organizing for independent contractors, for example, this option may be vulnerable too, because long-term collaboration is subject to drift and erosion over time.

More technological oriented suggestions are exploring the use of single technologies associated with artificial intelligence such as machine learning or blockchains for identifying health and safety risks as well as other labour standard violations. In a recent study Kurian and colleagues (2020) have examined to what extent algorithmic text analysis might be used to classify better common work accidents and their causes based on a large corpus of accident reporting texts. They show how accident diagnostics may be improved in terms of risk assessment and accident prevention later on. The benefit of this approach is that it may provide an interesting option to leverage the large amount of unused accident reports. However, the pitfall of this approach resides in its ex post orientation and its reporting focus. An obvious extension would be to use similar techniques with the big data from the underlying real production process which might be assessed instantaneously in order to actually prevent accidents from occurring. Inasmuch as this also involves the personal data of workers in the facilities this may require workers voice on how these data are assessed for other purposes.

Similarly, “smart” data base management tools based on the blockchain idea have been recently promoted as tools for ensuring compliance with labour standards along the supply chain (for one example). Embedded in a broader debate on how blockchain technology might be used for a more sustainable and decentralized mode of human governance (e.g. Dapp, 2019), the basic idea is that blockchain technology makes products carry the information about their origin with them which may also include information on how they were produced in each single step of the supply chain (Ponce del Castillo, 2020). However, the pitfalls of a blockchain approach towards labour standard compliance resides in the multiplexity and multidimensionality of business transactions’ quality, in contrast to electronic currencies for instance, as embedded in their societal contexts. For example, information about labour standard violations must somehow be fed into the system, and the truth values of this information might be controversial as there are unresolved power issues at the bottom of the pyramid as well along the further steps in global supply networks. In short: Who is in control of validating a textile worker’s complaint in Bangladesh about the violation of her right to organize collectively which she feeds into a database by using her mobile phone? The solutions suggested so far either assume actors who fill in accurate information and confirm these (even if it this potentially endangers their interest positions) or they suggest forward looking approaches in which credible actors are brought together who are already interested in lifting compliance with extant labour standards from the start. Hence, blockchains may operate for fostering labour standard compliance in contexts such as corporations voluntarily applying consumer labels. However, neither sanction-based compliance of violators nor moral evaluation, legal obligation and political controversy can be handed over to a neutral technology. Also, it remains highly unclear in a labour inspection context what it means that no intermediaries appear to be necessary. On the one hand, access barriers to technological infrastructures might be high because of prohibitive initial investments in equipment and know-how. This issue of articulation may be resolved somehow, for example by establishing digital forms of indirect worker representation. Still, then, the controversy remains around how violations are processed and negotiated and what the consequences are. Often, mere blocking of transaction partners would not do justice to complex local realities, at least if an improvement of working conditions on the ground is the intention.

In conclusion, digital labour inspection may become a real option in the future for strengthening compliance with labour standards in a digitalized world of work. Whether this potential can be realized, however, depends on improved technical solutions which avoid confusing virtual compliance with what is going on in real workplaces. Furthermore, a rejuvenation of labour inspection requires more organizing efforts, resources, and empowerment than just installing new software facilitating the documentation of labour standard violations and their consequences. The question is not so much about whether smart computers take over the control over moral judgment and legal sanctioning from humans, but whose rules are implemented in the algorithms and who is in charge of checking the installed systems for their proper functioning in practice; and, what resources, including those of a more traditional kind, are made available for that purpose.

References
Autor, D. H. (2015). Why are there still so many jobs? The history and future of workplace automation. Journal of Economic Perspectives, 29(1): 3-30.
Borjas, G.J. & Freeman, R. (2019). From immigrants to robots: The changing substitutes for workers. RSF: Russell Sage Foundation Journal of the Social Sciences, 5(5): 22-42.
Brynjolfsson, E. & McAfee, A. (2014). The second machine age. Work, progress, and prosperity in a time of brilliant technologies. Norton: New York.
Cappelli, P. & Keller, J.R. (2013). Classifying work in the new economy. Academy of Management Review, 38(4): 575-596.
Charles, J., Ferreras, I. & Lamine, A. (2020). A freelancers’ cooperative as a case of democratic institutional experimentation for better work: a case study of Smart-Belgium. Transfer, 26(2): 157-174.
Choudary, S. P. (2018). The architecture of digital labour platforms: Policy recommendations on platform design for worker well-being. ILO Future of work Research Paper #3. International Labour Office: Geneva.
Crouch, C. (2020). Will the gig economy prevail? Polity: Cambridge.
Culpepper, P.D. & Thelen, K. (2020). Are we all Amazon primed? Consumers and the politics of the platform power. Comparative Political Studies, 53(2): 288-318.
Dapp, M. (2019). Toward a sustainable circular economy powered by community-based incentive systems. In Treiblmaier, H. & Beck, R. (eds). Business transformation through blockchain. Vol II. Palgrave Macmillan. https://doi.org/10.1007/978-3-319-99058-3_6
De Stefano, V. (2020). Introduction: Automation, artificial intelligence and labour protection. Comparative Labor Law & Policy Journal, 41(1): 1-12.
De Stefano, V. & Aloisi, A. (2018). European legal framework for „digital labour platforms“. Luxembourg: Publications Office of the European Union.
Head, S. (2014). Mindless. Why smarter machines are making dumber humans. New York: Basic Books.
Healy, J., Nicholson, D. & Pekarek, A. (2017). Should we take the gig economy seriously? Labour & Industry: a journal of the social and economic relations of work, 27(3): 1-17.
Kuhn, K.M. & Maleki, A. (2017). Micro-entrepreneurs, dependent contractors, and instaserfs: Understanding online labor platform workforces. Academy of Management Perspectives, 31(3): 183-200.
Kurian, D., Sattari, F., Lefsrud, L. & Ma, Y. (2020). Using machine learning and keyword analysis to analyze incidents and reduce risks in oil sands operations. Safety Science, 130: https://doi.org/10.1016/j.ssci.2020.104873
Srnicek, N. (2017). Platform Capitalism. Cambridge: Cambridge University Press.
Upchurch, M. (2018). Robots and AI at work: the prospects for singularity. New Technology, Work and Employment, 33(3): 205-218.
United Nations Conference on Trade and Development-UNCTAD (2019). Digital Economy Report 2019. Value creation and capture. Implications for developing countries. United Nations: Geneva.
Vallas, S. & Schor, J.B. (2020). What do platforms do? Understanding the gig economy. Annual Review of Sociology, 46: 16.1-16.22.
Wood, A. J., Graham, M., Lehdonvirta, V. & Hjorth, I. (2019). Networked but commodified. The (dis)embeddedness of digital labor in the gig economy. Sociology, 53(5): 931-950.
Woodcock, J. & Graham, M. (2020). The gig economy. A critical introduction. Polity: Cambridge.
Plesner, U. & Husted, E. (2020). Digital organizing. Red Globe Press: London.
Ponce del Castillo, A. (2020). Blockchain in the world of work: hype or hope? Foresight Brief #9. European Trade Union Institute (ETUI): Brussels.
Silberman, M. & Johnston, H. (2020). Using GDPR to improve legal clarity and working conditions on digital labour platforms. ETUI Working Paper 2020.05. Brussels.