AI and the Transformation of Epistemic Regimes: the Problem of Hybrid Dependency

Abstract

The article examines how the pursuit of technological sovereignty in the field of artificial intelligence, instead of bringing true autonomy, generates a wide range of hybrid dependencies. These dependencies arise at the levels of computing infrastructure, data governance, AI system evaluation methodologies, and deployment infrastructure. It is shown that such dependencies can be comprehensively analyzed from both technical and philosophical-epistemic perspectives, as they directly affect the criteria of reliability, interpretation, and quality of AI-generated knowledge. Attention is paid to the phenomenon of epistemic security — the need to preserve a state’s ability to independently produce and interpret knowledge. The paper also analyzes the regulatory gap between the goals of state policy and the actual market consequences of AI implementation. The authors argue for moving away from the binary “domestic vs. foreign” opposition toward a multifactorial assessment of technologies, considering risks related to data, sanctions vulnerability, contextual applicability, and quality criteria.