Welcome to Switzerland, AlgorithmWatch!
Imagine you live in a Swiss city, and that you, and your whole family, are very excited that your child, who has just turned six years old, will start first-grade next summer. However, your home sits between two schools, and you don’t know which school your child will go to. Worse than that, you don’t get to decide which of these schools your child will go to because an automated decision-making (ADM) system could soon decide for you. Perhaps the two schools have very different reputations and, maybe the pupils come from very different social and ethnic backgrounds. You gossip with other parents about the relative merits of each school because, like them, you want the best for your child too. For some, ‘the best’, could mean the school which has a higher rate of pupils who go on to higher education, for others a good art education might be a priority, or maybe the popularity of the teachers could be a deciding factor. The opinions of parents will differ, as they always do when it comes to children. Several studies have shown that irrespective of a child’s background, segregation of children according to the social and ethnic background at school affects performance. A Swiss ADM system has been developed (although it has yet to be implemented) to assign pupils to schools in a way that mixes children from different backgrounds as much as possible. The system will also take into account the distance between home and school, the number of places and classrooms available, and the teaching personnel at each school. Theoretically, this mix of factors should help ensure an inclusive, equitable, high-quality education for all children. In addition, a diverse mix of social and ethnic backgrounds would also, arguably, help build social coherence and increase cultural diversity from a young age, which might prove useful for the future of your child in the globally connected world. On the face of it, all of this sounds very much like something most parents could get behind. However, some parents might not be happy handing this decision over to an algorithmic system. They might feel that this arrangement would put their child at a disadvantage. They may argue that their child’s education will suffer, that their grades would be lower, and, subsequently, that their higher education opportunities will be at risk. The use of ADM also raises political questions, including questions about the societal implications of using such algorithms in the education system. This is just one example of an automated decision-making system under development in Switzerland—there are many others—and how the implementation of these systems brings both opportunities and risks. As can be seen in the Swiss chapter of the Automating Society 2020 report, ADM systems already have an increasing impact on many parts of Swiss life. However, the somewhat unique federal structure of Switzerland means that the implementation of these ADM systems is subject to a diverse range of legal frameworks. This is where AlgorithmWatch comes in. As a research and advocacy organization, AlgorithmWatch sheds light on the social, legal, and ethical implications of ADM systems that affect society. AlgorithmWatch builds bridges between the private and the public sector on the one hand and with civil society organizations on the other. AlgorithmWatch is already well-known as an instigator of an ongoing dialogue surrounding ADM systems, and it continuously nurtures this conversation across Europe. With the launch of AlgorithmWatch Switzerland, we can now perform the same role in Switzerland. AlgorithmWatch Switzerland, supported by the Migros Pioneer Fund, officially started operations in September 2020, when our managing director, Dr. Anna Mätzener, joined the organization. AlgorithmWatch Switzerland is very much looking forward to engaging with more people, more organizations, and starting new and exciting projects that build upon our existing network in the country and expanding it over the coming months and years.