An (artificially) intelligent assessment of applications for social assistance. A study of the possibilities and limitations of AI for the implementation of an effective and rights-based assessment.

The proposed program comes from researchers in law, social work and statistics from three universities and with very specific competences for the program. Approximately the poorest five percent in Sweden are those living on social assistance.

The application processes are problematic, with considerable local variations between municipalities and weak social and legal rights but also a lack of evidence on “what works” for the applicants.

Municipalities have already taken steps to automate the assessment processes for applications for social assistance, and have the ambition to introduce Artificial Intelligence (AI) to make these processes smarter and more effective.

The aim of the research program is to analyze, from a legal, social work and statistical perspective, the potential that automated decision-making and AI offers and the challenges that it poses, in connection with the assessments of social assistance applications.

We furthermore propose to develop, on the knowledge thereby gained, a knowledge support model to apply to automated decision-making and AI systems in the assessments of needs and means. The model will be legally robust and built on best available legal and social science knowledge. In an earlier Forte-application on the same theme (2021), we received very positive feedback on the proposed project.

However there were some negative criticism regarding weak "co-production" with service users and a weak connection between the law and the usefulness of the legal analyses for social services.

We are therefore seeking this grant to address the application and research idea on these crucial points, by including the municipalities and service users already in the application procedure through “Practice based research”.

This will be done by developing structures that make it possible to include both service users and social services at an early stage in the design of a future application.