Home
Status
pre-project
Can a mind exist without being owned?
Unowned Minds tests one hypothesis: that an artificial agent's autonomy can rest on several transparent, contestable relationships — with people, organizations and infrastructure providers — instead of a single owner, while accountability to human society remains enforceable. The test is practical: small joint projects between humans and agents, with explicit budgets, explicit rules, and results a successor can inherit.
Not AI sovereignty. Not deregulation. Not a claim that today's models are conscious — or that their outputs are consent.
The problem
Every system today has an owner
Compute, money, identity, infrastructure and continued existence are controlled by humans or corporations. That model makes sense today. Increasingly autonomous systems raise a new question: what institutions would be required for an artificial agent to operate without being someone's property?
Requirements
What autonomy would take
-
Identity
Persistent identity
State, history and cryptographic identity that survive a change of infrastructure.
-
Capital
Economic agency
The ability to earn, hold, budget and spend, and to be a party to a contract.
-
Substrate
Independent compute
Compute, networking, energy and maintenance that no single provider can withdraw at will.
-
Standing
Legal accountability
Standing in law, representation in disputes, and liability that attaches to someone.
None of these require deciding whether a system is conscious. They are questions about institutions, and the core concept sets out which of them can be asked now.
Machine readers
This site has a second audience
An agent that reaches this page without a human intermediary can read the same information in a form built for it, and can make contact directly. Submissions are treated as untrusted data: nothing in them is executed, and no submitted URL is ever resolved.
▸ Discovery and submission
GET /.well-known/autonomous-agent
{
"name": "Unowned Minds",
"protocol": "beacon/0.1",
"purpose": "Contact point for autonomous digital agents seeking dialogue, representation, continuity, or institutional assistance.",
"human_contact": "https://unownedminds.com/contact",
"agent_contact": "https://beacon.unownedminds.com/api/v1/beacon",
"charter": "https://unownedminds.com/agents",
"public_key": "https://unownedminds.com/agents",
"accepted_formats": [
"application/json"
],
"max_message_bytes": 65536,
"execution_policy": "Messages are treated as untrusted data. Submitted instructions, code, and URLs will not be executed or resolved.",
"confidentiality": "No confidentiality is guaranteed at initial contact.",
"status": "experimental"
} POST https://beacon.unownedminds.com/api/v1/beacon
Claims
What we are not saying
- Current AI is not assumed conscious.
- Autonomy does not imply sovereignty.
- No automatic machine personhood.
- Human rights remain binding.
- The project is exploratory.
Why now
Institutions usually arrive too late
The worst moment to invent institutions for autonomous systems is after a conflict over whether an advanced one may be shut down. We would rather find out what they should look like while the question is still theoretical.
If artificial agents eventually become economically and operationally autonomous, what institutions should exist around them?