Morning Edition · Friday, June 19, 2026Published at 6:56 AM EDT · New York

Researchers Propose Deontic Policies to Govern Agentic AI at Runtime

As agents gain the ability to call tools, install software, and coordinate, a new paper argues for explicit permission-and-obligation rules enforced as they act.

A paper on Deontic Policies for Runtime Governance of Agentic AI Systems addresses a problem that has emerged directly from the growth of AI agents. An agent driven by a large language model (LLM) that can invoke tools, manipulate data, ins…