Research · Working Papers

The papers underneath MAR500.

Seven working papers establish the substrate, the canonicalisation methodology, the variance decomposition, and the empirical findings. Three are published to SSRN. Three are published to Zenodo with persistent DOIs. The seventh sits as a working draft pending temporal data from the next monthly seal.

Method note · Open access PDF

Decision Persistence: operating a large language model as a measurement instrument

The operating method behind MAR500: segregated roles, append-only decision records, pre-registered acceptance criteria, independent validation, and human release authority, mapped to the EU AI Act and DORA. Read it in the browser.

Open the PDF →
PDF · 18 pages · W. M. Collins · Meridian Autonomy B.V.
The Stationary Sea series
Part 1
The Stationary Sea (Part 1: Substrate Construction)
Measurement Instrument Validation for External Assessment of AI Governance in Regulated Financial Institutions
Zenodo · 10.5281/zenodo.19955199
Annex 1a
Canonicalisation Methodology, Variance Decomposition, and the Development Sequence
Annex to The Stationary Sea (Part 1)
Zenodo · 10.5281/zenodo.20182178
Part 2
The Stationary Sea (Part 2: The Long and Winding Road)
Substrate Identification Reproducibility, Two-Lane Canonicalisation, and the Multi-Institution Empirical Validation of the Hundreds-Not-Thousands Counter-Prior
Zenodo · 10.5281/zenodo.20185351
Topical findings on SSRN
Paper 1
The Coase Inversion
Topological Governance of Autonomous AI Agents in Regulated Financial Services
SSRN · 6470098
Paper 3
Credit Ratings Cannot Measure What They Cannot See
The case for an independent AI governance rating in financial services
SSRN · 6524438
Paper 4
Systemic Governance Risk in AI Agent Networks
Cross-institutional contagion and topological vulnerability
SSRN · 6535599
Paper 2
Cascade Without Containment
AI agent contagion under DORA and the EU AI Act · Working paper pending temporal data
Working paper
Citation & access

Zenodo publications carry persistent DOIs and are archived for long-term reference. SSRN papers are open access. For citation queries, replication packages, or institutional access to the underlying substrate, contact [email protected].