Short videos showing algorithmic (un)fairness
Collaboration with AlgorithmWatch Switzerland on the exhibit "(Un)fair Algorithms"
More information: Project in the digital guide of the exhibition
Algorithms are increasingly used to make fundamental decisions in our lives: Will I get a loan? How high is my insurance premium? Am I entitled to welfare benefits?
These systems are often highly efficient. They have been trained with data sets from many people and, as a result, have learned to recognize patterns in these data sets. Those patterns are used, for example, to look for indications of the degree of risk that a particular individual will not repay a loan. In this context, it is not only important that the algorithms work well – in other words, that they do what they are supposed to – but also that decisions made based on them are fair.
AlgorithmTV: Wer kriegt den Kredit?
Even algorithms that, at first glance, make seemingly less vital decisions have an impact on our lives and how we live together as a society: How do social media algorithms decide which content to show me? Which ads do I see in my browser? Is Google Maps directing drivers through my neighborhood because of a construction site? Which bus stops does the algorithm suggest for the new transport network? Does it consider the need for connections to the local post office as well as fast links to the city center?
AlgorithmTV: Wer darf zum Bewerbungsgespräch?
The question is whether these systems are fair and reduce social disparities or whether, on the contrary, they maintain or even reinforce such inequality.
AlgorithmTV: Wem helfen wir aus der Arbeitslosigkeit?
Our four short videos draw on real events and that could easily be imagined to unfold in a similar way in many countries. They show the various aspects of inequality that need to be discussed when using algorithm-based decision-making systems. At the moment, they are only available in German.