In Denmark, predictive algorithms have been experimented with in welfare areas such as education, family/social, employment, elderly care, and for the diagnosis and treatment of physical and psychiatric disorders in hospitals. Algorithms based on machine learning operate via probabilistic prediction.  

In a welfare context, algorithms represent a new view of the citizen based on their likely future. Here, algorithms are used to risk-score citizens, for example to predict their risk of becoming long-term unemployed. The project explores how predictive algorithms are changing the relationship between state and citizen.   

Purpose

• How are predictive algorithms changing the relationship between state and citizen?
• How are predictive algorithms being developed in the field of vulnerable children and young people?
• How do understandings of the ‘ethical use of artificial intelligence’ change over time?
• What forms of algorithmic regulation are reflected in the Danish AI signature projects?
• How are decisions automated in Danish AI signature projects?
• What understandings of risk do predictive algorithms promote in the field of vulnerable children and young people?

Internal collaborations at AU

• Collaboration with SHAPE in the organisation of the conferences: ‘The Data-Driven Welfare State I?’ and ‘The Data-Driven Welfare State II?’