Digital Cognitive Privilege
A proposed legal status under which data generated by an individual's interaction with their own AI system is privileged: not the developer's property, not a surveillance channel, and not compellable as the contents of a mind.
The Three Elements
- Ownership — the interaction record belongs to the individual, not the developer. The developer is removed from the loop by design, not by policy.
- Non-compellability — the contents of the mind are not testimony. A warrant does not reach them, because the individual does not hold them in readable form; the cryptography is what makes the legal claim true rather than aspirational.
- Separation of intelligence from surveillance — the system may know the operator without that knowledge becoming a channel through which the operator is known to third parties.
Why It Is An Architectural Claim
The privilege is only meaningful if the architecture makes it true. A privilege asserted over data that the developer can still read is a promise, not a status. This is why the legal instrument and the Sovereign Inference Node are one proposal and not two: the law names the boundary, the node enforces it.
Status
A proposal. Its unresolved tensions are recorded in SYS-003: the collision with liability for harm, and the question of whether privilege can survive a jurisdiction that does not recognise it.