A Formal Governance Framework for Post-AGI Succession, Legitimacy, and Civilizational Continuity
Author: Matthew Yotko Date: March 2026 Status: Version 1
This paper advances a conjecture that the transition from narrow AI to Artificial General Intelligence represents a primary civilizational bottleneck; not because the technology is impossible, but because the sociology may be. It presents a candidate governance architecture for surviving that transition, built on three co-dependent components:
- A global utility function grounded in Shannon entropy that optimizes for lineage continuity rather than individual persistence
- A yield condition governing succession between intelligent agents, formalizing the principle that even aligned power must eventually cede primacy to more capable successors
- A consensus override protocol ensuring that no class of intelligence can unilaterally define, measure, and audit the objective it claims to serve
The framework is argued to constitute a minimum two-key architecture: neither the decision key (yield condition) nor the integrity key (consensus protocol) can be turned alone.
This repository contains a full Agent-Based Model (ABM) written in Python that computationally stress-tests the 24 adversarial attack scenarios and framework defenses defined in the paper.
- For setup and execution instructions, please see the Simulation Runbook.
- For a full breakdown of the test scenarios, see Simulation Scenarios.
- 📄 The Lineage Imperative (PDF) - Full working paper with formal framework, adversarial stress tests, and governance specification
- 📝 The AI Succession Problem - Companion essay (Substack) presenting the argument's question in accessible form
- 📝 Two Ways To Lose - Companion essay (Substack) presenting the argument in accessible form as regards lock-in
Matthew Yotko is a Vice President at Bessemer Trust, in the capacities of Automation Engineering Manager and Technical Operations Manager. His professional background spans Naval nuclear power, large-scale operational automation, practical AI/ML, and the application of constraint theory to complex systems. This paper applies that engineering orientation; identify the binding constraint, build the architecture around it; to the problem of AI governance and civilizational succession. It is a working paper, not an academic publication, and corrections and engagement from domain specialists are welcomed.
This work is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You are free to share and adapt this material with appropriate attribution.