Human agency
Consequential choices remain attributable to people or an existing authorised system.
Core principles
The system is designed around six principles that continue to apply as capabilities change.
Consequential choices remain attributable to people or an existing authorised system.
One Canonical Person authority prevents competing identity truths and cross-person leakage.
Built is not Verified. Verified is not Live. Status moves only on direct evidence.
Purpose, consent, minimisation, isolation and deletion are designed with the feature, not added after it.
Personalisation, intelligence and commercial objectives cannot silently weaken Safety.
Duplicate, interruption, restart and unknown-outcome behaviour are treated as normal engineering cases.