Values
Pillar 3. A citizen captures their this-versus-that rules — what matters more, what is ethical, how a situation should be handled — as a structured value system. An AI can then apply those rules to real situations and produce recommendations consistent with the citizen's own ethics.
This is the pillar that turns subjective values into executable logic.
What a value system contains
Rules, read top-down. Each has a when (the conflict it governs), a prefer (which side wins), an
optional unless (the exception that flips it), a weight, and a rationale.
- id: r1
when: "individual liberty conflicts with collective safety"
prefer: "liberty"
unless: "imminent, specific, proven harm"
weight: 0.8
rationale: "Rights are the default; the bar to override them is high and evidentiary."
Principles — unconditional statements that are not this-versus-that.
Priorities — a ranked fallback for a conflict no explicit rule covers.
A partial value system is a valid value system
Everything except a name and an id has a sane default. A citizen who has written one rule has a working value system. You are not asked to axiomatise your ethics before the software will talk to you.
Policy packets
A value system travels. You can publish yours as a policy packet to the movement's social network, or import somebody else's whose judgment you trust — cached locally, attributed to its author, and yours to edit.
Publishing requires a real identity. Reading, writing, and applying your own never does.
Pages in this area
- Write down your values — how to start, and what makes a good rule.
- Policy packets — publishing and importing.
- Where this stops — the honest limits.
Where to go next
- Decisions — the area before this one.
- Evidence — the area after this one.
- Open the app — do this for real, on your own seats.