BlogAI Economics
AI Economics9 min read

AI Agents Won't Crash the Economy. Bad Governance Might.

An expert analysis of the AI science and economics behind the Citrini Research report on agentic AI.

By Richard Ewing·

The Macroeconomic View of Agents

Agentic AI—systems designed to take autonomous action—represents a fundamental shift in the variable cost of intelligence. For the first time, organizations can scale complex labor without scaling human headcount linearly.

The Liability Gradient and Macro Loops

However, an autonomous agent without a rigorous governance boundary represents uncontrolled liability. As agents begin to interact with other agents at high frequency, we witness macro regression loops: recursive errors caused by Agentic Drift where probabilistic models lose the thread of their original intent.

To prevent catastrophic failure, companies must deploy an Exogram Action Admissibility Protocol (EAAP) layer that deterministically checks the state and authorization of an agent before it commits an operation to a system of record.


For detailed EAAP specifications, see our Documentation Center. Original article hosted at Built In.

Like this analysis?

Get the weekly engineering economics briefing — one email, every Monday.

Subscribe Free →

More in AI Economics

Published Work

This article expands on ideas from my published work in CIO.com, Built In, Mind the Product, and HackerNoon. View published articles →

📊

Richard Ewing

The Product Economist — Quantifying engineering economics for technology leaders, PE firms, and boards.