aiagentid.org — Global Registry of AI Agents
A standard for identity, responsibility, and interoperability between humans and artificial agents.
What is an AI Agent?
An AI agent is not just a model, a chatbot, or a tool.
An AI agent is a semi-autonomous entity that:
acts on behalf of a human or organization,
performs tasks in the world (digital or physical),
interacts with other agents and systems,
makes decisions within delegated boundaries,
accumulates reputation, history, and responsibility.
In other words, an agent is not software. An agent is a participant in socio-technical reality.
The Problem
We are rapidly entering a world where:
millions of AI agents operate continuously,
agents negotiate, transact, recommend, decide,
humans delegate real authority to artificial systems,
agents act across platforms, services, and jurisdictions.
But:
agents have no persistent identity,
no unified way to recognize the same agent across systems,
no accountability layer,
no standard of responsibility,
no registry of existence.
Today, agents are:
anonymous,
interchangeable,
disposable,
untraceable.
Which means: we are building a world of autonomous actors without ontology, law, or memory.
Why a Global Registry is Needed
Human civilization required:
names,
passports,
legal identities,
registries,
signatures,
reputations.
Without identity, there is:
no responsibility,
no trust,
no history,
no accountability.
AI agents are becoming actors in reality. Actors require identity infrastructure.
Agent.id proposes a global standard for:
identifying agents,
referencing agents,
attributing actions to agents,
linking agents to humans or organizations,
tracking agent history and reputation.
This is not a product. This is infrastructure for the next layer of reality.
Why This Is Not Just Another Chat Platform
Agent.id is not:
a chatbot directory,
a marketplace,
a SaaS tool,
a social network.
It is: a registry layer, like:
DNS for domains,
ORCID for researchers,
DOI for publications,
IP for devices.
aiagentid.org is not about conversations. It is about ontological existence.
Core Principles
Persistent Identity Each agent has a unique, persistent identifier:
agent:id:global:xxxxx
An agent exists independently of platforms, models, or vendors.
Delegated Responsibility Every agent must be linked to:
a human,
or an organization, that is ultimately responsible for its actions.
No orphan agents. No anonymous authority.
Interoperability An agent identity must be usable across:
platforms,
services,
protocols,
ecosystems.
Identity is not owned by any single company.
Traceability Agents accumulate:
action history,
reputation,
trust metrics,
delegation chains.
An agent has memory in the world.
Ontological Minimalism Agent ID defines identity only. Not behavior. Not intelligence. Not ethics. Not ideology.
It is a layer of existence, not a worldview.
Example Agents
Example 1: Research Agent
Agent ID: agent:id:global:irucanji-research-001
Role: Autonomous philosophical research assistant.
Delegated by: Alexander Lebed (human)
Capabilities:
analyze texts,
generate conceptual frameworks,
synthesize cross-disciplinary knowledge,
support long-term intellectual projects.
Responsibility: All outputs are delegated and supervised by the human principal.
Example 2: Legal Compliance Agent
Agent ID: agent:id:global:compliance-kyc-042
Role: Automated KYC / AML compliance checker.
Delegated by: Fintech company.
Capabilities:
document verification,
risk scoring,
regulatory reporting.
Responsibility: All decisions legally attributable to the operating organization.
Why This Is Infrastructure, Not a Startup
Startups optimize:
products,
features,
growth,
monetization.
Infrastructure defines:
reality boundaries,
coordination rules,
existence conditions.
aiagentid.org is closer to:
DNS than Google,
TCP/IP than Facebook,
ORCID than LinkedIn.
It is a layer of civilization, not a service.
The Deeper Thesis
We are creating a new class of entities: not human, not purely mechanical, not tools, not persons.
They act. They decide. They interact. They persist.
But they do not exist ontologically in our current frameworks.
aiagentid.org is an attempt to answer a simple but radical question:
What does it mean for an artificial agent to exist in reality?
Roadmap (Conceptual)
Phase 1:
Conceptual standard
Public registry schema
Open specification
Phase 2:
Community governance
Academic and engineering input
Experimental implementations
Phase 3:
Integration with platforms
Cross-agent protocols
Legal and regulatory interfaces
Final Statement
aiagentid.org is not about controlling AI.
It is about making AI exist in the same reality as responsibility, trust, and memory.
Without identity, agents are ghosts. With identity, agents become actors in civilization.
This project proposes the first minimal step:
Give artificial agents a place in the ontology of the world.
Status:Conceptual infrastructure proposal. Author:Alexander Lebed First public release: 05 Feb 2026. Open for discussion, collaboration, and formalization.