Upplevda förutsättningar för en framtida implementering av AI som beslutsfattare och eXplainable AI : En kartläggande fallstudie på en svensk kommun baserad på Strategic Alignment Model’s fyra domäner

Sammanfattning

Problem: Swedish municipalities are expected to digitize and streamline their operations with the help of Artificial Intelligence (AI). One way is by using AI as an automated decision maker (AI ADM), which poses strategic, legal and ethical challenges. For decisions made with the help of AI to be understandable, transparent and legally secure, new ways of working are required, where the technical documentation process eXplainable AI (XAI) can serve as a supporting tool. However, there is a lack of knowledge about what conditions are required for AI ADM and XAI to be implemented together in the Swedish municipal sector. Purpose: The purpose is, based on the theoretical framework Strategic Alignment Model's (SAM) four domains, to map a Swedish municipality's conditions for a future potential implementation of AI ADM and XAI. Method: The study has had an interpretivist and constructionist viewpoint, inductive approach with deductive elements and a cross-sectional case study. The study has been conducted in a Swedish municipal context, where data collection is based on semi-structured interviews and document studies, which have been analyzed through thematic analysis in a multi qualitative design. Conclusion: The study concludes that the Swedish municipality has very complex conditions, which directly affect their future possibilities to implement AI ADM and XAI. By mapping these given circumstances based on SAM's four domains, a comprehensive description has indicated that the Swedish municipality's current business strategic, IT strategic, organizational and IS infrastructure conditions, constitute obstacles for a future implementation of AI ADM and XAI.