Research Repository
Browse my research papers on AI governance and model reliability. Each entry includes an abstract and a direct link to read the full PDF. Select a paper title or its “Read the full paper” link to open the document.
Adaptive AI Governance as a Lifecycle Control System: From Static Compliance to Continuous Oversight
ABSTRACT
This white paper introduces Adaptive AI Governance, an approach that reframes AI governance as a closed-loop lifecycle control system rather than a static compliance exercise. The framework integrates upstream Boundary Governance for dataset certification, supplier integrity, epistemic traceability, and representational diversity enforcement with the Adaptive Lifecycle AI Governance Framework (ALAGF), which orchestrates continuous monitoring through MIDCOT for drift detection, BME Auditor Pro for behavioral evaluation, SymPrompt+ for runtime alignment enforcement, and LGSB for structured stress testing. The closed-loop feedback architecture ensures that operational signals from all subsystems propagate bidirectionally, enabling real-time corrective action and upstream revalidation when governance thresholds are breached. The approach is explicitly aligned with the NIST AI Risk Management Framework and ISO/IEC 42001, demonstrating that integrated instrumentation and continuous evidence generation are prerequisites for substantive standards compliance.
Dale Rutherford · 2026 · 18 pages
ABSTRACT
Model autophagy (the recursive ingestion of synthetic or contaminated data through iterative retraining cycles) threatens the epistemic integrity of large language models at scale. This paper presents a four-layer causal architecture integrating session-level confirmation bias, corpus-level data contamination, generation-level parameter drift, and civilizational-scale epistemic consequences. Formalized through 21 recurrence equations governing corpus integrity I(n), bias B(n), and quality Q(n), the model predicts exponential decay trajectories with phase transitions around I(t) = 0.5. Validated through controlled GPT-2 recursive retraining experiments (10 generations, S(t) 0.10 to 0.80), Phase 3a (5 tracks, 55 observations) confirms exponential decay (alpha = 1.93, R-squared = 0.98, floor = 0.468). Phase 3b (5 tracks, 55 observations) demonstrates that G2+G3 governance (w_p = 0.60) partially reverses decay: I(10) = 0.894 versus baseline 0.489 (U = 0.0, p = 0.004). Calibrated BRF = 0.115 yields FIF*BRF = 0.179, satisfying the stability condition. These results provide the first empirical demonstration that provenance-based governance can arrest and partially reverse epistemic decay in recursive AI training.
Dale Rutherford · 2026 · 12 pages
Governing BME Amplification in Large Language Models: A Standards-Aligned Framework for LLM Data Lifecycle Risk Detection and Ethical AI Governance
DOCTORAL DISSERTATION
This dissertation develops a standards-aligned framework for detecting and mitigating bias, misinformation, and error across the large language model lifecycle. It introduces the BME Metric Suite and connects behavioral measurement, information-quality monitoring, and governance interventions through the ALAGF, MIDCOT, and SymPrompt frameworks.
Dale A. Rutherford · University of Arkansas at Little Rock · 2026 · 313 pages
Deposited version DOI: 10.5281/zenodo.23884501
Interaction-Induced Knowledge Narrowing Risk in LLM Systems
PUBLISHED JOURNAL ARTICLE · REVIEW
This paper identifies Interaction-Induced Knowledge Narrowing (IIKN): a lifecycle risk in which bounded context windows and iterative interaction constrain the alternatives surfaced by an LLM, even when its responses remain accurate and fluent. It distinguishes unobserved exclusion from conventional model errors and proposes governance controls, including epistemic scope declarations, breadth-before-depth interaction, and coverage confirmation checkpoints.
Dale Rutherford and Ningning Wu · Discover Artificial Intelligence · Volume 6, article 1153 · September 17, 2026 · 23 pages
Read the published article · DOI: 10.1007/s44163-026-02101-6 →
Open access · CC BY-NC-ND 4.0