This essay aims to identify which public values motivate and legitimise the implementation and increased use of artificial intelligence (AI) in Swedish municipalities.

It also examines which public values risk being deprioritised when AI is implemented, with particular focus on democracy, transparency and accountability.

The study is conducted through a qualitative text analysis of municipal policy documents from four Swedish municipalities that differ in size and in their degree of AI maturity. The material is complemented by reports and guidelines from national authorities and other relevant actors. The theoretical framework for this essay is Public Value Management.

Mark Moore’s strategic triangle is used to highlight how different public values are formulated, prioritised or downplayed in the material.

The results show that efficiency and operational capacity are recurring motives for AI implementation, and are often linked to demographic change, resource constraints and increasing service demands. Democratic values such as right to appeal, transparency and accountability are in principle recognised as important, but are treated in more general terms and given less concrete content. This suggests that these democratic values risk being undermined and deprioritised in favour of operational efficiency goals. Overall, the study indicates that AI in municipal policy documents is primarily legitimised through effifiency and operational capacity, while democratic dimensions of public value are pushed to the background. This raises questions about how municipalities can balance effeciency demands with democratic legitimacy.