Skip to content

Research

The Terrene Foundation publishes research on the governance of autonomous systems. This includes analysis of failure modes in existing AI governance approaches, design patterns for human-AI collaboration, and the theoretical foundations underlying the CARE, EATP, and CO specifications.

  • Failure analysis: Why current approaches to AI governance fail in practice: regulatory capture, compliance theater, trust asymmetries, and the gap between policy and implementation.
  • Trust architecture: How trust can be made verifiable at the transaction level, and what properties a trust protocol must have to remain meaningful at scale.
  • Organizational design: How organizations can structure themselves to maintain genuine oversight of autonomous systems without creating bottlenecks that eliminate the benefits of autonomy.
  • Constitutional governance: The role of constitutional structures in preventing governance decay, capture, and drift in open-source foundations.

Constraint Theater: Autonomous Governance Beyond Compliance is the foundational thesis of the Terrene Foundation. It models the governance problem for autonomous systems using Laffont-Tirole regulatory frameworks, proves conditions under which organizational separation (Trust Plane / Execution Plane) is optimal, and derives the CARE, EATP, and CO specifications from formal propositions. The thesis is in preparation for academic publication.

The three Foundation specifications are published under CC BY 4.0:

  • CARE (Collaborative Autonomous Reflective Enterprise): the governance framework
  • EATP (Enterprise Agent Trust Protocol): the trust verification protocol
  • CO (Cognitive Orchestration): the methodology for human-AI collaboration

See Standards for specification details.