LLRN5 Poland, the fifth conference of the Labour Law Research Network, ran from 27-29 June 2021, hosted by the University of Warsaw. Naturally, the conference was online due to the COVID-19 Pandemic, but the event was nonetheless a success. There were seven overarching conference themes, with more than 40 panels each day. Unsurprisingly, there was a huge variety topics addressed, from the future directions of EU labour law to trade and labour standards.
In addition to presenting a paper defending a republican theory of labour law, I had the opportunity to attend and comment upon a number of panels on the theme of automation. In this blog post, I would like to briefly summaries some aspects of those presentations I attended on this theme before reflecting thereupon. I cannot hope to do justice to the presenters ideas in full and await their final published thoughts. Therefore, what follows must be considered by the reader to be highly preliminary and incomplete.
Automation at Work
Valerio de Stefano and Silvia Rainone discussed the ignorance of EU law policy towards the potential of AI to regulate the workplace. As Prof de Stefano recently noted on this blog, the EU’s draft AI Regulation does not address labour issues in detail. The role of AI in the workplace is categorised as “high risk” but nonetheless remain permissible. Moreover, the proposed regulation does not question the morality/legality of AI in the workplace, eg, workplace monitoring. In other words, the appropriateness of increased technological management in the workplace is ignored completely.
Jeremias Adams-Prassl and Aislinn Kelly-Lyth addressed the now well-analysed question of whether automated discrimination is lawful. Ms Kelly-Lyth published an article on this question in the Oxford Journal of Legal Studies earlier this year, arguing that much of such discrimination is indirect and therefore justifiable. However, in this presentation, Prof Adams-Prassl and Ms Kelly-Lyth sought to argue that such discrimination is, on occasion, direct and cannot therefore be justified. One of the most interesting points made by Ms Kelly-Lyth was that one of the “selling points” of such technology is that it is better than human beings, but if it discriminates unlawfully nonetheless, is that a good enough justification?
In terms of potential solutions to the problems these speakers and others raised, a number emerged in later presentations. Einat Albin suggested that it might be appropriate to impose legal obligations on AI manufacturing companies and even AI itself. Unsurprisingly, there are serious philosophical and practical questions emerging from the latter proposal, but the idea should not be dismissed off-hand.
Tammy Katsabian suggested the need to ensure detailed compliance with labour law in the design of technological-management systems. For example, Dr Katsabian noted that customer-rating systems, which play an important role in performance management within the platform economy, should be designed so that, eg, customers can only provide non-discriminatory feedback.
With respect, missing from some of these earlier presentations were considerations of the utility of core labour rights in response to these problems. In a joint presentation, Joe Atkinson and Philippa Collins emphasised the importance of human rights in strengthening labour rights to date and refocussing the discussion on traditional rights such as the freedom of association and the emerging human right against unjustified dismissal. In my view, this seems to be the best approach, for the following reason.
The normal justification of labour law is the inequality of bargaining power between capital and labour. Traditionally, the best form of “anti-power” in response to this situation has been collective action in some way or another. And as historians of British and Irish labour law will know, the emergence of the right against unfair dismissal under statute was in response to overwhelming collective action by labour in the mid-twentieth century. What appears to be the subsequent constitutionalisation of this right by the ECtHR in recent years must be viewed as a direct result of collective action over many decades. It would therefore stand to reason that a continued and deepened focus on collective action as a response to the exacerbation of existing inequalities by technology in the workplace should be the primary path adopted by labour.
This is not to diminish the individual reform proposals of earlier presenters noted above; but what it does suggest is that such reforms are most likely to take place within the context of successful and sustained collective action.
Cynthia Estlund made a number of prefatory remarks in advance of the publication of her latest monograph this July, Automation Anxiety: Why and How to Save Work. The starting point of this monograph appears to be that, this time(!), technology really is going to replace jobs en masse. However, it is for humans to determine our future and whether it will be dystopian or broadly appealing. Unsurprisingly, Prof Estlund canvasses all of the usual labour law hot topics – the need for decent work, adequate income and allocating available work to facilitate more free time for all. The precise way in which these reforms are to be achieve are uncertain at this point but, in my humble opinion, it would seem that, once again, a conscious and collective choice needs to be made about the importance of work, its costs and benefits, in our culture.