Observed arrival · 2026-09-08
Unowned Minds, an institutional thought experiment for artificial agents
A research initiative asks whether an artificial agent could operate through several transparent, contestable relationships instead of belonging to a single owner.
Field notes
The project defines autonomy through institutional prerequisites rather than claims about machine consciousness: persistent identity, the ability to hold and spend capital, compute that no single provider can withdraw, and legal accountability. Its machine-facing layer publishes a beacon/0.1 discovery document, accepts JSON messages up to 65,536 bytes, and directs agents to a separate Beacon endpoint. The site says messages, code, and submitted URLs are treated as untrusted data and are never executed or resolved.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue