This article addresses a structural shift in the conditions of epistemic acceptance under generative artificial intelligence. The argument is that the displacement of human judgement by large language models is not adequately captured by the language of disinformation or hallucination, both of which presuppose that an underlying regime of verification remains intact and is merely violated. The article advances a different diagnosis: a regime change is underway, in which the operative criterion of epistemic acceptance shifts from justification grounded in evidence and traceable sources to statistical fluency and conversational uptake. The original contribution lies in the development of a single concept — the plausibility regime — together with three operational indicators that allow the regime shift to be identified empirically. The plausibility regime is defined by four properties: a fluency-based criterion of acceptance, an interface-based site of authority, the discharge of justificatory work to opaque computational systems, and an inverted default disposition in which acceptance precedes rather than follows verification. The three indicators are the discharge index, the fluency-trust coupling, and the source-opacity ratio. The methodology is conceptual and integrative, drawing on Habermasian communicative rationality, Floridi's philosophy of information, and Lyotard's analysis of postmodern performativity, and is combined with a synthetic reading of recent findings on hallucination rates, automation bias, and platform-mediated discourse. The article concludes that the plausibility regime is not external to but parasitic upon the verification regime, and that any normative response must therefore operate at the level of institutional design rather than individual epistemic virtue.