How will AI agents change agency, labor, markets, organizations, responsibility, verification, and the distribution of economic value?
← Back to publications
BOOK DETAIL · Civilization Science Series · 04
AI Agent Economics
The Economic Architecture of Agentic Civilization
An economic framework for AI agents as delegated action systems that allocate tasks, use tools, exchange information, affect prices, reshape organizations, and transmit risk under human authorization.
01
Research overview
AI agents create a new layer of delegated economic agency; analysis must connect objectives, authority, execution, verification, liability, reward, and system stability.
Key concepts
01Delegated economic actor02Human–agent economic unit03Agent task and service markets04Verification and mechanism design05Data, compute, identity, and trust06Responsibility, liability, and systemic risk
02
Research architecture
I · Agent boundary
Capability, autonomy, delegation, identity, authorization, and termination.
II · Production and labor
Human–agent units, task decomposition, organizations, productivity, and displacement.
III · Markets
Discovery, pricing, contracting, data, compute, payments, and platform power.
IV · Verification
Evidence of action, output quality, contribution, audit, and dispute resolution.
V · Risk and governance
Goal errors, concentration, manipulation, finance, cascades, accountability, and rights.
VI · Civilizational context
Cooperation, trust, entropy, value distribution, and long-term human agency.
03
Method and evidence program
Open questions for review
- Q1
Which rights and responsibilities must remain non-delegable to AI agents?
- Q2
How should an agent’s contribution be attributed among user, developer, model, data, platform, and institution?
- Q3
Which market structures preserve competition, human agency, and system stability under high automation?
04
Academic status and limits
This book is a published foundational research output. Its concepts, equations, classifications, and causal claims remain open to independent review, operationalization, counterexamples, replication, and revision.
The book does not treat AI agents as persons by default and does not provide investment, legal, token, or automated-governance advice.Intended readers
Economists, AI researchers, organization designers, mechanism specialists, policy researchers, platform builders, and human–agent systems teams.
05
Publication record
Suggested citation
Hu, Aurion. (2026). AI Agent Economics. Civilization Science Series. Amazon Kindle edition. ASIN B0H8LNY5TV.
Errata and version record
No edition-specific public errata file is currently linked. Corrections should identify the edition, location, original text, proposed correction, reason, date, and effect on conclusions.