A New Type of Intellectual Interaction: Dialogue Through a "Disobedient" Mediator - IA
Year: 2026
Type: preprint
Abstract: This article describes a new type of intellectual interaction emerging when two researchers use generative AI models as mediators for creating scientific concepts. Unlike traditional dialogue (human-human) and dialogue with AI (human-machine), a three-level structure arises: researcher A → AI → researcher B → AI, with feedback loops and the effect of AI "disobedience" generating unexpected results. The metaphor of a "puppet duel" is proposed: two opponents engage in an intellectual duel, controlling puppet marionettes (AI models) that are not entirely obedient. It is shown that it is precisely the unpredictability of AI that creates new turns in the dialogue and allows going beyond the limits of each participant's thinking. The features of the new type of communication are analyzed: distributed agency, asynchronicity, accelerated iterativity, and removal of personal vulnerability. The article discusses how this changes the process of creating scientific knowledge and what challenges it poses for the philosophy of science, epistemology, and research practices. The article is a meta-reflection on the series of works on SFOС, Principle G, the Projection Coherence Index (PCI), and the verification program for ontological models. A real dialogue between the author and his colleague, during which these concepts were created, is used as an illustration. The article concludes with a call for researchers to recognize their participation in a new epistemic practice and to develop new norms of scientific communication.
Language: English
Subfield: History and Philosophy of Science
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7212254180
Cites:
Cited by:
Related to:
