Proof-of-Impact + Human Authorization Infrastructure for AltmanAI
Humanity leads. Intelligence follows.
The paihi-framework repository contains the core definitions, governance model, reference patterns, and tooling specifications for AltmanAI's P-A-I-H-I (Proof-of-Impact, Alignment, Integrity, Humanity, Impact) system.
This is the foundational layer that ensures every AltmanAI product and internal action maintains human authority and produces verifiable positive impact.
- Formal definition of the P-A-I-H-I framework
- Human Authorization patterns and boundaries
- Impact verification and logging standards
- Reference implementations and integration guides
- Governance rules and safety constraints
- Templates for Proof-of-Impact Ledger entries
In active development. This repository is being established as part of AltmanAI's July 2026 infrastructure build.
We are currently focused on:
- Core framework documentation
- Security and governance baseline
- Clear contribution and usage guidelines
Note: a working reference implementation of the P-A-I-H-I scoring logic (PAIHIScorer, Python, tests passing) already exists in human-first-ai. This repo will absorb and formalize that implementation as the canonical spec matures — the two are not intended to remain permanently separate.
- Human authorization is required for all high-impact actions.
- Every significant action must produce verifiable Proof-of-Impact.
- AI supports human judgment — it does not replace it.
- Security and documentation are non-negotiable infrastructure.
See the docs/ directory (coming soon) for:
- Full P-A-I-H-I Framework Specification
- Human Authorization Implementation Guide
- Proof-of-Impact Ledger Template and Rules
This repository follows AltmanAI's security standards. See SECURITY.md for vulnerability reporting and responsible disclosure.
Apache License 2.0 — see LICENSE file.
We welcome aligned contributions once the initial framework is stabilized. Please review the governance rules and open an issue first for significant changes.
This framework powers all AltmanAI products, including DailyPilot and future tools. It is the foundation of our commitment to human-first, verifiable AI.
Maintained by: AltmanAI by Altman Family Group LLC Contact: Admin@blakealtman.com (for security issues and all other inquiries) Web: https://altmanai.tech Last updated: July 2026