Large Language Models as Revolutionary Media Objects: Rethinking Relationality, Agency, and Knowledge in the Age of Generative AI

The rise of generative AI, like large language models (LLMs), highlights a profound shift in the contemporary media landscape. LLMs produce deep mediatization effects, being part of the emerging generative infrastructure of contemporary digital capitalism. This paper conceptualizes LLMs as revolutionary media objects, drawing on theories from actor-network theory, object-oriented ontology, and critical AI studies to argue that their generative capacities enable new forms of non-human forms of machine semiosis, synthetic subjectivity, and recursive epistemologies. These generative models decenter anthropocentric frameworks and showcase a need for reevaluation of the relationship between humans, objects, and meaning and knowledge production via the phenomena of machine semiosis and digital habitus. Through a systematic literature review and theoretical synthesis, this work introduces concepts such as synthetic relationality, generative ecologies, and agentic co-entanglement to analyze how LLMs reconfigure sociality, power, and epistemology. The paper offers a novel framework for understanding the role of generative AI in reshaping contemporary platform infrastructures, with significant effects toward central concepts like agency, sociality, and knowledge production. This paper contributes to ongoing discussions within media theory and critical studies on AI about the role of LLMs in contemporary society.