Epistemic Provenance Mesh Protocol (EPMP): A Diagnostic Instrument in Human-AI Communication
Year: 2026
Type: article
Abstract: The Epistemic Provenance Mesh Protocol (EPMP) is a diagnostic framework for verifying the genuine human origin of epistemic contribution in human-AI dialogue — distinguishing it from AI-scaffolded coherence, where the thinking has been delegated to the model rather than generated through the human's own epistemic architecture. It operates through two mutually constitutive layers — the Upstream Governance Framework (UGF), eleven cross-cutting competencies constituting the epistemic architecture that determines whether genuine human contribution is present, and the Epistemic Imprint Diagnostic (EID), which reads that contribution's presence, absence, and morphology directly from the dialogue trace. The EPMP requires no external instrumentation and no proxy measures — it operates exclusively through native observables already present within existing LLM architecture, rendering it immediately executable across closed or open weights LLM interfaces operating with a defined upstream training objective. This publication represents the first version of the Epistemic Provenance Mesh Protocol (EPMP) full working paper. Included within this version are the high-resolution figures corresponding to the three iterations of the proof of concept programme executed by the author. Version 2 of the full working paper coming soon. The iteration transcripts (complete, timestamped dialogue exports) are available upon request at [email protected]. During the proof-of-concept phase, they are not provided in open access to preserve independent instantiation of the protocol. Early exposure to the original dialogue traces risks biasing replication through pattern imitation rather than generating the epistemic conditions the framework is designed to detect. The transcripts constitute the primary temporal evidence of the diagnostic (including collapse timing, hysteresis, and arc-level structure) and are therefore best engaged within the framework that renders them interpretable. Access is provided upon request to support methodologically grounded evaluation while maintaining the integrity of independent testing. Feedback from EPMP protocol testing is welcomed directly by the author. Ongoing engagement with the framework contributes to its development and will inform future licensing decisions. DOI: https://doi.org/10.5281/zenodo.19197150 License: CC BY-NC-ND 4.0 Architectural Notice: The Epistemic Provenance Mesh Protocol (EPMP) is designed as an integrated architecture consisting of the Upstream Governance Framework (UGF) and the Epistemic Imprint Diagnostic (EID). Please read the full licensing and architectural notice situated on the cover of the working paper.
Author: Balanovski, Igor Alexei
Language: English
Subfield: Safety Research
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7130933779
FWCI:
Cites:
Cited by:
Related to: