Entry
Governance-as-a-Service: A Multi-Agent Framework for AI System Compliance and Policy Enforcement
Suyash Gaurav, Jukka Heikkonen, Jatin Chaudhary
Proposes a runtime governance layer that regulates agent outputs without altering model internals; introduces a Trust Factor and three intervention types (coercive, normative, adaptive); evaluated on open-source models for content/financial scenarios.
·governance-as-a-service ·multi-agent ·policy enforcement ·trust factor ·runtime
- AI Organizations are More Effective but Less Aligned than Individual AgentsApril 11, 2026 · arXiv
- Institutional AI: Governing LLM Collusion in Multi-Agent Cournot Markets via Public Governance GraphsJanuary 16, 2026 · arXiv
- CooperBench: Why Coding Agents Cannot be Your Teammates YetJanuary 19, 2026 · arXiv