Being's (g)host in the machine. How AI is turning ontology inter-actional

Abstract Large Language Models (LLMs) are not merely powerful tools for language manipulation, they are philosophical provocations. While current discourse treats them as knowledge systems or statistical approximators, their behavior signals something deeper: a tectonic shift in our assumptions about being, ‘meaning’, coherence, and even process itself. It would not be the first time in history that a new technology, like microscope, telescope, pump, clock or computer, changes our worldview. This white paper argues that LLMs expose the limitations of the classical ontological paradigm, grounded in separate entities, fixed types, and truth as correspondence with their internal essence. Instead, they compel us to confront an emergent, inter-actional kind of ontology, where ‘meaning’ arises as interacted consequences in context, coherence is negotiated, and identity is transient. This is not just a technical or epistemological disruption, but an ontological inversion: from what is to what happens. We explore how this shift unsettles disciplines from AI and epistemology to ethics and system design, and why a new philosophical framework, one that is adaptive, participatory, and temporally sensitive, may now be necessary. Interestingly, this paper was submitted to Philosphy & Technology and declined by Editor in Chief Luciano Floridi