The public sector is embracing algorithmic infrastructures to offer innovative digital services to citizens. This research project assesses perspectives and practices among experts, providers, and citizens for an inclusive and democratic ecosystem of future AI governance. The European Commission's proposed Artificial Intelligence Act has raised debates on the implications of this regulatory framework across the EU. We create an annotated corpus of texts on public AI services and use Critical Discourse Analysis and Natural Language Processing to examine diverse values, narratives, and positions. We conduct case studies of public AI services to highlight practices and challenges for AI-based technologies in the public sector. We explore and evaluate citizen perspectives, agency, and imaginaries on digital citizenship and algorithmic literacy for democratization of public services to design frameworks for stakeholder participation in AI governance in Finnish society and the EU.