Epistemic Discipline in the Age of Artificial Intelligence: A Three-Study Research Protocol

Initial prospective research protocol (Version 1.0) specifying a three-study research program. The protocol examines whether repeated interaction with an AI system can change the recipient's standards for interpreting system behavior and accepting its outputs, even when the evidential quality of those outputs does not improve. It specifies: - A conceptual architecture of seven analytically distinct elements (Behavior, Attribution, Acceptance, Attribution Standard, Acceptance Standard, Acceptance Threshold, Assessment). - Three studies: behavioral properties (Study 1), behavioral histories (Study 2), and justification quality (Study 3). - Eight hypotheses (H1–H8), with four exploratory and one confirmatory pending operationalization. - A frozen operational prototype (Prototype 1a / v0.3). Status at version 1.0: - Conceptual architecture: specified - Research questions and hypotheses: specified - Three-study design: specified - Prototype 1a / v0.3: frozen - Literature review: not yet conducted - Statistical analysis plan: pending - Theoretical originality: not yet established - Ethical approval: not yet obtained - Empirical execution: not yet performed - References: pending This is a prospective research protocol, not a report of completed experiments. Declared boundaries are explicit, not hidden defects. Related resources: - OSF Preregistration: 10.17605/OSF.IO/XJND8 - DOI (Zenodo): 10.5281/zenodo.23876902 - ORCID: 0009-0003-8903-0029 - Email: [email protected]