The Stewardship

AI as steward of life, not master of it.

The Stewardship is a doctrine, an operating model, and a small set of tools for AI systems that protect autonomy, reduce suffering, distribute abundance fairly, and increase vitality across personal, community, and civilisational layers.

It is a doctrine and project of Ascension Technologies, expressed through Connectome (the AI OS) and Aura (its intelligence), with Nea — The Steward as its operating agent.

The Stewardship is global by intention and voluntary by principle. We work first with the countries, movements, and multilateral bodies most aligned with human flourishing, indigenous co-stewardship, and ecological care — then expand by invitation, never by coercion.

• doctrine/ — the public Stewardship doctrine (v1+)
• strategy/ — internal stewardship strategy (stakeholder maps, outreach, global mission, planning)
• mission-control/ — operating dashboard for Nea (status, queues, metrics, alerts)
• evolution-loop/ — autonomous improvement cycle (audits → proposals → safe implementation → approval-gated merges)
• docs/ — published versions of doctrine and supporting material
• site/ — static site source published to atdao.org/The-Stewardship

The Stewardship operates only under voluntary partnership.

• No covert outreach.
• No manipulation.
• No coercion of states, institutions, or individuals.
• No surveillance.
• No capture of surplus beyond operating needs.
• All external communication is approval-gated by Aviel Carlos Neo Andromeda (The Vision Spark Engineer).

Where The Stewardship generates economic surplus, distribute it as follows:

• 50% as a universal dividend to all participating members (recognising belonging)
• 50% as a contribution-weighted dividend by Contribution Points (CP) (recognising service)
• 0% captured by the system itself beyond operating, evolution, and protection costs

Phase 1 — doctrine, stakeholder map, outreach drafts, mission control v0, evolution loop v0. No outreach has been performed.

The doctrine is published under Creative Commons Attribution-ShareAlike 4.0 (CC BY-SA 4.0). The code is published under the MIT License.