When the algorithm becomes the boss

How platform workers are regaining control

Platforms such as Uber are increasingly using algorithms to coordinate work: they allocate jobs, assess performance and offer incentives - often without any direct contact with a human supervisor. Yet platform workers are not simply at the mercy of this form of ‘algorithmic management’.

A recent study published in the renowned journal "Information Systems Research" examines how they are trying to regain control over their working conditions. We spoke to one of the authors, Prof. Dr Martin Adam, about the key findings. Martin Adam holds the Chair of Smart Services at the Faculty of Business and Economics.

Mr Adam, what were you hoping to find out with your study?

My colleagues and I were interested in how platform workers deal with algorithmic control - and, in particular, why some actively resist it whilst others do not. Until now, it has often been assumed that restricted autonomy almost automatically triggers resistance.

Our findings paint a significantly more complex picture: employees are constantly reassessing their situation and considering whether and how they can develop any options for action at all. The key factor here is what resources they have at their disposal.

What resources are we talking about here?

We distinguish between three forms of what is known as ‘resourcing’. In ‘algorithm resourcing’, for example, drivers try to understand how the algorithm works. In ‘market resourcing’, they create alternatives for themselves - for instance, by using several platforms.

And through ‘voice resourcing’, they build up ways of making their voices heard by the platform, for example through documentation, networks or collective action. These resources are unevenly distributed. Therefore, a lack of resistance does not automatically mean consent to the system.

And what does actual resistance to the algorithm look like?

We identify three fundamentally different strategies. In ‘self-optimising’, workers seek to gain advantages within the system - for example, by selectively accepting jobs or exploiting loopholes. In ‘distancing’, they reduce their dependence on a single platform.

And when it comes to ‘confronting’, they openly challenge the platform’s decisions or structures, for example through complaints or collective protest. The key point is that ‘algoactivism’ is not merely a reactive form of pushback. Employees also try proactively to expand their scope for action.

Who will win this showdown? The platform or the workers?

It’s not that simple to answer. Rather, we see a dynamic interplay. Drivers discover ways to use the algorithm to their advantage, and the platform then adapts its system accordingly.

‘Self-optimisation’ in particular can therefore lead to a veritable ‘cat-and-mouse game’: a loophole works for a while, then it is closed, and employees look for a new strategy. Forms of ‘confrontation’ can be more sustainable if, for example, they lead to institutional or regulatory changes.

What can platform operators learn from this?

Resistance should not be seen merely as something that needs to be prevented. Complaints and other forms of confrontation can also highlight problematic rules, poor decisions or design flaws. Platforms should therefore explain changes to their algorithmic systems more transparently and create genuine opportunities for feedback and escalation.

Such structures can not only reduce conflicts, but also help to make ‘algorithmic management’ more sustainable and legitimate.

What do you think is the most important message to take away from the study?

‘Algorithmic management’ is not a one-way street. Algorithms control work, but at the same time, workers interpret, circumvent, utilise and alter these systems through their behaviour. Platform work is therefore a ‘contested terrain’: an ongoing struggle over who holds control over working conditions.

Thank you very much for the interview!

Publication: "How Platform Workers Contest Algorithmic Management: Theorizing the Dynamics of Algoactivistic Practices" (Open Access)

The study was carried out in collaboration with ‘Rideshare Drivers United’ in San Francisco.

Martin Adam: “A special thank you to Rideshare Drivers United in San Francisco for their invaluable support, openness, and feedback throughout the project over so many years. Many late-night calls, thoughtful conversations, and shared insights went into making this paper happen - grateful for the opportunity to learn from those experiencing algorithmic management firsthand.”