AI Society Theory: Re-imagining Castells’ Network Society in the Algorithmic World

The rise of artificial intelligence (AI) challenges existing frameworks in media and communication studies, revealing the limits of classical theories and Castells’ network society in explaining contemporary social transformations. While Castells (2010) conceptualized society as structured through networks, flows, and informationalism, the proliferation of AI systems introduces a new layer of mediation, where algorithms autonomously generate, filter, and govern information, culture, and economic activity. This paper proposes AI Society Theory as a conceptual framework to address this gap. Building on the foundations of network theory, we reconceptualize the space of flows as the space of AI and extend informationalism into algorithmic informationalism, highlighting how AI systems actively shape knowledge, visibility, and value creation. We examine the formation of layered AI communities- from users of specific platforms, to developer collectives, to platform-based subcultures- demonstrating that AI infrastructures are sites of social interaction, identity formation, and cultural production.Through a critical review of classical media theory, the network society, platform society, surveillance capitalism, and essential AI studies, we identify a gap: existing frameworks either neglect algorithmic autonomy or remain centered on human agency. AI Society Theory addresses this by positioning AI as both infrastructure and social actor, offering a lens to analyze how communication, power, and culture are reorganized in algorithmic societies. By articulating the theoretical contours of AI-mediated social life, this study lays the foundation for future empirical and conceptual research, providing a robust framework to understand how AI is transforming societies, platforms, and communication in the twenty-first century.