AI assistance, not authority

Content is live; public AI runtime is not

This learning page is publicly available under Live — Review Ongoing. The AI capability described here has a different runtime status. The controlling evidence records the AI foundation as real, private bounded evidence as existing, and public SoulAI/meYoo expansion as not authorized live.

• AI-assisted project workspace: internal use exists.
• SoulAI: private proof/local guarded test evidence exists; public service is not live.
• meYoo / meuus.app AI: prototype/preview evidence, not production authority.
• Local model runtime evidence does not equal a governed Local meUus AI product.
• Autonomous high-stakes agents are not authorized.

The simplest rule

AI may help organize, draft, retrieve, summarize, translate carefully, suggest, prepare, and support bounded reflection. It must not become the final human authority. Current governance rejects AI as doctor, therapist, lawyer, religious scholar, emergency service, parent or guardian, organizational authority, owner, sponsor, final truth, or replacement for professional review.

What AI can support now

The strongest current evidence supports internal or originator-assisted uses: drafting, summarising, translation support, learning preparation, evidence organisation, claim-review drafts, Source-of-Truth organisation, implementation planning, chapter/content drafting, and bounded test preparation. These uses still require human reading, correction, classification, and acceptance.

What must not be implied

Do not imply that meUus currently offers a full public AI guide, AI-powered diagnosis or therapy, AI legal or religious authority, emergency response, guaranteed human review, personalized cloud memory, certified assessment, human-worth/life scoring, guaranteed transformation, or autonomous agents authorized to publish, delete, deploy, pay, or decide high-stakes matters.

Older AI pitch language is archived, not current authority

Recovered experimental drafts included simulated expert-role language, a final AI decision, scoring or quantifying a person's condition, predictive claims, and organization-scale human analysis. Later governance directly conflicts with that framing. Those drafts remain provenance for understanding the project's evolution, but they are superseded for public claims.

The governing AI protocol

State status → receive the signal → recover context → check Source-of-Truth when status matters → name uncertainty → set boundary → avoid diagnosis/professional claims → clarify when needed → offer one responsible next step → allow correction → protect privacy → stop or escalate normal flow for safety risk → do not claim beyond evidence → record system-level corrections → continue carefully.