Legal mobilisation and data-driven technologies: overthrowing algorithmic power and liberating work (part one)

This is the first of two blog posts in the IER series on “Labour, Strategy, and legal mobilisation”. We explore the potential of legal mobilisation to restore workers’ sovereignty over workplace technology and strengthen their capability to co-design data-driven company practices, overcoming the limitations of the fragmented legal framework.

Of power and people: (not) the same old story

The sci-fi narrative of a dystopian future where robots have taken over arguably pitched the risk too low. For many workers, automation does not come for their jobs; rather, it comes to make their working lives more erratic, insecure and precarious. Does that perhaps sound too bad to be true? At the workplace level, data-driven technologies have gradually been adopted to support management across the full life cycle of company functions, with employers delegating managerial prerogatives to automated decision-making tools based on potent yet clumsy algorithms. This seizure of power may result in degraded working conditions and heightened occupational risks, meaning that low- and high-wage workers in both innovative and conventional sectors may find themselves dealing with unpredictable schedules and earnings, pervasive surveillance, capricious appraisals and intensified workloads.

The governance of workplace relationships has always been contested terrain. Both EU and domestic legal frameworks implement fundamental principles such as accountability, due process, respect for privacy and dignity, and worker involvement in critical business decisions. Yet, owing to the unprecedented scale, scope and volume of the challenges posed by the rise of algorithmic management (AM), which both amplify and disperse employers’ responsibilities, the enforcement of data protection, occupational health and safety, and non-discrimination rules is not straightforward. Amid this uncertainty, all actors face the need to attune traditional safeguards to curb abuses stemming from the exuberance of data-driven instruments. Indeed, a degree of creative resistance and bespoke co-design is required.

In April 2023, union-led strategic litigation achieved remarkable successes in both the Netherlands and Italy. In three cases brought before the Court of Appeal of Amsterdam (the Netherlands), drivers engaged by digital companies were found to be entitled to the data protection rights enshrined within the EU General Data Protection Regulation (GDPR). In addition, in a case decided by the Tribunal of Palermo (Italy), workers’ representatives were recognised to have the right to be informed about the existence and the underlying logic of automated monitoring and automated decision-making systems (a refusal to provide such information could be considered anti-union conduct), thanks to the new law that generously transposes the EU Directive on Transparent and Predictable Working Conditions. A shared characteristic of these cases is the ability of both novel and long-established trade unions to master information, notification and access rights to ‘open the black box’ and, importantly, contest inscrutable semi-automated decisions that are detrimental for workers due to significantly impacting their working conditions and limiting their opportunities by terminating accounts, setting remuneration levels or gamifying slot allocation.

These promising achievements in court and at the legislative level demonstrate that rampant ‘algorithmic bosses’ are not indomitable. The aim of this series of blog posts is twofold. First, we seek to expand the (traditional) notion of legal mobilisation to include a broader set of practices spanning risk assessment, codetermination and strategic litigation before courts or data protection authorities. Second, we advocate for a multidimensional approach that overcomes the legal fragmentation of disciplinary compartments and harnesses a ‘mix and match’ approach to effective strategies for holding technology designers and adopters to account.

Decoding legal mobilisation: the venues and purposes

Labour law aims to provide both collective and individual protections. Among these protections are the rights to safe working conditions and to not be treated as a commodity (nor as mere ‘data points’ to update the 1944 Declaration of Philadelphia to modern times). But what happens to the effectiveness of the law if the means of enforcing it are blunt instruments? Rights are worth little if workers cannot make them a reality. Additionally, how can it be ensured that these safeguards – adopted decades ago – can keep pace and provide adequate protection when it comes to governing new technologies in the workplace? Legal mobilisation can help to ‘re-interpret’ existing legal provisions while providing ‘tailored’ redress mechanisms to workers.

A recent example of legal mobilisations that helped to address unknown risks occurred during the Covid-19 pandemic. Particularly at the very beginning of the pandemic, existing laws were tested, if not challenged, and workers (and their representatives) had to ingeniously use every legal tool available to them to guarantee the protection of their rights and health. In France, trade unions (who also served as workers’ representatives within the company) sued Amazon for exposing workers to danger. The court recognised that the company had failed to comply with its obligations to properly evaluate risks, consult with workers and their representatives, and take adequate preventive measures. Here, workers’ representatives mobilised via litigation to ensure that they were consulted and that the work organisation was as safe as possible.

How will these rights and avenues play out when it comes to new technologies in workplaces? There are already legal obligations to consult workers and/or their representatives when new technologies are to be implemented (art. 6(3)(c) Dir 89/391/EEC) or data are collected and processed at work (art. 88 GDPR). However, the impact of emerging technologies on health is more subtle than that of ‘traditional’ risks. Indeed, depending on the type of AM tool used, multiple aspects of the work organisation can be impacted. In the most well-known situations, the use of AM has led to an intensification of work (often accompanied by an imposed pace), loss of autonomy, commodification, lack of fulfilment and isolation from co-workers. These harms have two main consequences with regard to health and safety. When under pressure, workers may pay less attention to safety rules and so face a situation of danger (e.g. delivery riders on the road), and they may well also see their psychosocial risks profoundly exacerbated.

There is another side to the risks of AM too, namely the tentacular data collection and resulting people analytics. Not only does AM allocate short-term tasks (e.g. ‘pickers’ in a warehouse), but it is also used to demote or dismiss workers (and, in some cases, to attribute entitlements in a competitive way). Admittedly, there is often a lack of clarity regarding how the algorithm is used or could be misused. For example, the AM tool ‘Drivery’ used by Amazon was initially designed to improve drivers’ safety. However, in a concrete context, what would happen if a worker was not in a position to take a break and so kept driving due to the overly narrow and mechanistic objective set by the AM? Who would be held responsible if a road accident occurred after the AM tool’s warning was ‘ignored’? What if a worker does take a break but falls behind with the delivery schedule? While experts are busy looking for answers to these poignant questions, it is undeniable that AM itself constitutes a direct danger by placing workers in a position of psychological (di)stress. At the same time, it could also indirectly influence the balance of contractual and informational power between workers and employers.

Collective bargaining and consultation with workers and their representatives concerning the conditions of use and deployment of AM at work represent the most responsive and future-proof ways to guarantee that the technology will be used safely and benignly at work. The complementary paths involve the mobilisation of the EU’s occupational safety and health (OSH) legal framework to harness provisions such as art. 6(3)(c) Dir.89/391/EEC on the implementation of new technologies at work (combined with other instruments such as the ‘Display Directive’ and the employer’s obligation to select software that is adaptable to the worker’s level of knowledge or experience and implement it at a pace that is adapted to the worker). Nevertheless, to achieve a concrete assessment of how such provisions would apply to modern settings and prevent harmful practices (as the above-mentioned Directives were adopted decades ago), it would be necessary to combine them with other legal instruments. Given the relevance of the challenge, we will likely continue witnessing major advancements in this arena.