The nature of science in the age of generative artificial intelligence: Epistemic authority, materiality, and a synergetic response

Abstract

Generative artificial intelligence has entered science classrooms faster than the epistemological questions it raises have been addressed. This study is an attempt to reveal how the knowledge-producing character of text-based conversational systems diverges from institutional science, and what that divergence implies for teaching the nature of science in secondary and undergraduate science. The two are compared through the family resemblance approach across aims and values, methods, practices, knowledge, and social-institutional systems, and the procedure linking literature to judgement is reported. The comparison is then reframed through synergetic: science is an open system in continuous evidential exchange with material reality, whereas the model reorganizes a corpus closed to new evidence at inference. A coupled-system model follows: the classroom verification norm is the order parameter, the teacher-set verification requirement the control parameter, and materiality the boundary condition. Five propositions are derived, the last a path-dependent threshold prediction that can falsify the model.