Gerard Rinse OOSTERWIJK

FEPS Expert Consultant and Political advisor to the S&D VP Alex SALIBA

Does regulation make Europe uncompetitive? Or is the well-protected European social model a blessing when we need to innovate in the interplay between workers and AI systems? Specific rules on algorithmic management and AI in the workplace can boost productivity if the EU empowers workers and trade union representatives to steer this tech revolution and co-create better work processes. 

So far, the second Von der Leyen Commission has chosen one central theme for its work: the competitiveness of the European economy. This was expected, given the Draghi and Letta reports. It also reflects the reality that European industry is heavily impacted by the current geopolitical situation. However, the EU has neither matched the US Inflation Reduction Act nor provided state support equivalent to what China provides to its electric car producers. 

The harsh reality is that the union’s financial power is limited, so the advice for deregulation to ease the burden on businesses has more traction than recommendations to invest strategically in sectors and specific industries. We will have to see the impact of the competitiveness compass launched by the Commission and what will be in the omnibus package aimed at repealing some burdensome rules but potentially rolling back Europe’s green and social ambitions. 

While reducing the amount of rules is the current buzz, it is interesting to examine what the European Commission has committed to regulating. One hint is in the mission letter to the Socialist vice-president Roxana Mînzatu, who is tasked to develop an algorithmic management initiative. A FEPS study found that about three-quarters of businesses use one or more algorithmic management tools; a more recent OECD study even found that 79 per cent of European companies use these innovative tools to steer and control their workers. 

The world of work is changing profoundly because of these AI-powered software tools, which adversely affect workers, leading to less motivation, loss of trust, higher workload and more stress. That is why more specific rules are required, in addition to the AI Act that already indicated that the workplace is a high-risk use case for AI. The automation of decisions, which previously were in the hands of managers, with algorithmic governance tools makes it necessary for labour laws to be updated. We can think of specific rules on the transparency of and worker influence over algorithmic systems, the human in command of decisions taken that affect workers and putting a stop to the deterioration of occupational health and safety risks due to algorithmic management. These rules already apply to platform workers under the new Platform Work Directive but should be applied to all workers.

As FEPS, we hosted newly awarded Nobel laureate Daron Acemoglu to reflect on this matter. One thing I learned is that this algorithmic management revolution of our work does not have to mean bad news for labour and trade unions. Specific forms of automatisation made possible through algorithms, what Acemoglu calls ‘so-so automation’, self-checkout in grocery stores for example, will merely replace workers but not increase productivity. Real innovation using the potential of the algorithmic tools will require knowledge of the production process. This is precisely why firms need their workers and organised labour to work with them for better results. The firms that understand this might get ahead of their competition, but there will be a push to cut costs and get short-term gains from the so-so automation. Therefore, the European legislator should not shy away from setting rules that empower workers to innovate with AI, and, at the same time, protect them from the more destructive aspects of algorithmic management.

Our research shows that workers’ influence and also transparency mitigate the adverse effects caused by algorithmic management. However, these aspects of co-creation can even be seen as a precondition for making the introduction of AI effective and attaining the potential productivity rise. That is why the European Social Model, with its consultation and co-determination, could give us a competitive edge over other economic models like the Anglo-Saxon model where trade unions are weaker, and workers’ rights are less of a concern. 

To make something of this opportunity, the last thing we need is to roll back on worker rights or introduce a new 28th regime of company law that could be used to circumvent certain protections. We need to enable trade unions to co-create, together with management and the workers they represent, the right conditions to use the potential of algorithmic tools while upholding or even improving working conditions and sharing the benefits of productivity gains. However, to do this, management must also control the tools they deploy, which, often, they do not. The tools are black boxes for them too, usually purchased from Silicon Valley big tech firms that do not consider the role of trade unions in the processes of their European clients. This is where the discussion of algorithmic management links in with another big topic: Europe’s tech dependency. 

Currently, the EU depends for 80 per cent of its tech on foreign (non-European) providers. This lack of autonomy is felt in the workplace when algorithmic systems are deployed and developed in a different context. European-developed algorithmic management solutions should be high on the agenda of EU Commissioner Henna Virkkunen’s tech sovereignty goals. One could envisage an ecosystem of open-source algorithmic tools that can be deployed and adjusted in coordination with workers. This is not just because it is good for their well-being but, first and foremost, as a strategy to use our competitive advantage to innovate. A highly skilled workforce that, due to its workers’ rights, can co-create the new work processes that will bring about the gains of the AI revolution for businesses and society. 

An algorithmic Management Directive is badly needed because it will give the framework for trade unions to negotiate. Even better, the limits we put on the big tech surveillance capitalist tools we are all forced to use create the space for tech alternatives to be developed based on European values like workplace democracy.

Photo credit: Shutterstock / Pixels Hunter