🫧 Technical Foundations
— So that technology does not break Nagi’s centerless structures
Nagi’s technology is not for managing society through one screen, one ledger, or one AI.
It is a foundation that helps people and communities continue choosing their own lives, memories, and culture: helping, connecting, and explaining only when needed.
1. Five principles
- Polycentricity: Do not depend on one operator, one AI, or one data foundation.
- Choice: Preserve the options to use AI, consult only people, move to another AI, leave no record, or exit midway.
- Minimal data: Do not handle data whose purpose, necessity, and duration cannot be explained.
- Explanation and challenge: Attach grounds, uncertainty, alternatives, and a route for challenge to each proposal.
- Portability: People can inspect, export, and move their records, settings, and scope of delegation to another route.
2. Technical layers
| Layer | Role | What it protects |
|---|---|---|
| Personal layer | The person manages memory, settings, consent, and scope of delegation | deletion, portability, suspension |
| Community and organizational layer | Services for culture, education, care, and collaboration | local autonomy and context |
| Interoperability layer | Exchange only the minimum necessary information with a purpose and time limit | consent and purpose limitation |
| Public layer | Translation, accessibility, and shared knowledge | access for everyone |
| Audit and remedy layer | Explanation, complaints, incidents, bias, and objections | independence and review |
Every layer retains an alternative route that can be used with people alone.
3. Handling records
A record does not always increase trust. Records are separated by purpose.
- Personal memory: Managed by the individual and portable or deletable.
- Shared records: Purpose- and time-limited, with personal information minimized.
- Cultural archives: Accompanied by an intention to inherit and by context, with a chosen scope of publication.
- Public audit records: Retain only what is necessary for institutional accountability.
The freedom to decide what not to retain takes priority over recording everything.
4. Boundaries for AI and automation
AI may help translate and summarize materials, support accessibility, organize issues, offer counter-hypotheses, and draft explanations.
AI does not score human value or emotion, make final decisions about basic life or rights, or create long-term memory or emotion inference without consent.
AI design follows The Nagi AI Charter, Freedom, Dissent, and Exit, Memory, Consent, and the Right to Forget, and The Ethics of Social Systems.
5. Audit and re-breathing
Technology can fail. Models can be biased. Operators are not complete.
Nagi’s technology includes a route for requesting explanation, independent audit, human review, a means to halt high-impact functions, reporting of incidents and bias, and periodic data reduction.
Re-breathing is not only an update that adds functions. It is also an update that reduces unnecessary surveillance, excessive automation, and drift toward centralization.
6. Order of implementation
| Stage | What is built | Condition of success |
|---|---|---|
| Phase 0 | Human-only routes and optional conversational AI | the same support remains available without AI |
| Phase 1 | Personal vaults and screens for deletion, portability, and consent | people can handle their own records |
| Phase 2 | Interoperability among multiple routes | no closure within one provider |
| Phase 3 | Audit, challenge, and incident response | errors can be stopped and repaired |
| Phase 4 | Local and cultural public-AI commons | autonomy and diversity remain protected |
Technology is not the center that completes Nagi.
It is a foundation that quietly steps back so people can live without taking one another’s breath away.
→ Nagi AI Charter
→ The Ethics of Social Systems
→ Japanese reference edition