This paper proposes AgenticRei, a deontic policy framework that enables runtime governance, constraints, and deontic reasoning for autonomous multi-agent systems. By decoupling policy enforcement from the LLM via an external logic engine, it establishes a foundational security and compliance architecture for agentic AI.
Short on time? Watch the video overview of this paper.
Listen on: Spotify · Apple Podcasts · Amazon Music · YouTube
Follow the show: YouTube · Medium · X · LinkedIn · Bluesky · Threads · Instagram · TikTok · Facebook · Telegram