Algorithms in child welfare service - opportunities and challenges

This workshop invites to engage participants’ reflections on opportunities and challenges with algorithms in child welfare services. The case is a research project developing a predictive risk modelling to support social workers decision-making when they must assess children who are at risk of child abuse or neglect. The purpose with the project called "Notifications in focus" is to investigate the effect of this. The workshop will be introduced with a short presentation with the project and the tool. We would like to start a broad discussion of opportunities and challenges (e.g., in relation to ethics, legislation, data, implementation, trust, skills and involvement of families). All participants will be asked to use a digital platform to engage active participation in collaboration.
Ensuring children's protection and their best interests is a central challenge in social work. These decisions must also be made in a short span of time, as otherwise there may be major consequences for the children. Developments in international social work with children, young people and parents are most often at a crossroads with complex decisions where the risk of not thriving must be assessed. Since the 1980s, there has been a focus on developing various tools for making risk assessments to support social works decision-making.
New technology such as algorithms can enhance the opportunities to work with statistical data (already existing data from the municipality). Preliminary findings from the research suggest that social workers find that algorithms can work as a decision support tool even though it cannot stand alone.
Other research has found that algorithm can cause trouble in relation with working the families and that the interest in preventing child abuse or neglect simultaneously contributes to discrimination and create more inequality. The Danish projects with algorithms are either closed or have changed direction due to a lack of clarity in the legislation.