The question of who actually rules whom is rarely easy to answer. This is also the case with algorithms. The problem is addressed by the sociology of algorithmic rulership.
Algorithms are becoming increasingly important in an increasingly digitalised world. They suggest our purchases, recommend music to us or determine routes, and thus influence our actions in the world in a very tangible manner.
But what are algorithms exactly? According to a popular definition, just like a recipe for cooking, they describe what should happen step by step, so that the desired result can be achieved. However, in the opinion of sociologist Professor Robert Seyfert, this falls far short of the mark: "It is not the case that algorithms simply execute one step of a 'recipe' at a time with cool, objective precision and perfect reliability. Rather, sociology shows that in practice, algorithms are always closely interwoven with their technical and non-technical environments." Seyfert is investigating precisely how they do this and what ideas of rulership develop from it, together with doctoral researcher Jan Groos in a research project on the rulership of algorithms.
"With rulership, we do not primarily mean governance, but every form of control over individual or collective action," explained the social scientist, who has been working in Kiel since April 2021. Often enough, such control processes take place in everyday life without even being noticed by those affected. The so-called "nudging" is directly aimed at changing behaviour through impulses that are almost unnoticed or at least not perceived as disturbing: the fly in the urinal encourages men to urinate more precisely, and food served on small plates makes hungry people feel full faster. Amazon and the like have long since also used the process in digital form by evaluating personal preferences.
We have the right to know what data is being processed, but it would be much more interesting to know what conclusions are being drawn from it.
The algorithms are also quite obviously at work in motor vehicles, from an increasing number of assistance systems right through to completely autonomous driving. "The question is, who is actually assisting whom?" said Robert Seyfert to describe the associated shifts in the relationship between people and technology. "Does the car help a person to maintain a constant speed, for example? Or does the person help the car if the assistance system loses its orientation because the corresponding map is missing?"
Within the joint research project, Jan Groos investigates such questions and also significantly more complex ideas of future rulership using algorithmic technologies. For example, he specifically looks at blockchain technology. Blockchains are essentially huge databases that are not centrally located in one place, but distributed in a network. The data stored in the blockchain is publicly accessible, so that any person with sufficient knowledge and the appropriate equipment can potentially participate in the process of verifying the integrity of the stored data, and attach new blocks to the chain. However, the term "decentralised" is "frequently used as a marketing term and often heavily ideologically charged," said Groos critically. He considers the claim that power structures within blockchain-based networks are "decentralised" to be highly questionable. And even more questionable is the assumption that blockchain-based technologies could be used to somehow automatically decentralise (political) power in other areas of society.
Meanwhile, Robert Seyfert has shown in empirical studies that the often critically-viewed high-frequency trading, with its transactions generated algorithmically in the blink of an eye, is not simply a result of technological progress, but rather a product of political regulations. According to his findings, this shows that such regulations can also help give rise to algorithms.
Similarly with data protection, rulership over the algorithms also does not seem to function entirely satisfactorily. According to Robert Seyfert, the EU General Data Protection Regulation does actually create more transparency about algorithms that are used on individual persons. However, it has a major weakness: "We have the right to know what data is being processed, but it would be much more interesting to know what conclusions are being drawn from it."
The research also offers highly interesting material on a theoretical level. Among other things, it is currently being discussed whether algorithms can perhaps be used for contemporary forms of planned economy and even for overcoming capitalism. Ironically, this would result in the fulfilment of the old dream of socialism with a human face – of all things, thanks to the soulless digital technology.
Author: Martin Geist
More in-depth information
Both Robert Seyfert and Jan Groos produce podcasts with discussions on the subject: in "Der Streit" (the dispute), Robert Seyfert and Andre Armbruster discuss new sociological literature. In "Future Histories", Jan Groos tries to "expand our vision of the future" with different guests.