governance: establish human-first AI release and accountability controls

[altmanAI]

Objective

Implement the next phase of ALT-001 by converting AltmanAI's human-first principles into enforceable, reusable GitHub engineering controls.

Closes part of #8.

Changes

• Adds .github/CODEOWNERS for attributable review ownership across governance, security, trust, templates, and profile surfaces.
• Adds an AI-system change/evaluation issue form requiring human authority, privacy, security, evaluation, limitations, rollback, and P.A.I.H.I. analysis.
• Adds AI_SYSTEM_CARD_TEMPLATE.md for models, agents, AI-assisted features, and tool-using workflows.
• Adds RELEASE_GOVERNANCE.md with risk-based release classes, universal gates, evidence requirements, staged rollout, release blockers, emergency-change controls, and retirement requirements.

Evidence and validation

• Branch is four commits ahead of main and zero commits behind.
• Only four new governance files are introduced.
• No production code, secrets, dependencies, permissions, or external services are changed.
• Rollback is a clean revert of this pull request.

Security and privacy

This change strengthens security and privacy review requirements. It introduces no credentials, data processing, runtime permissions, or third-party dependencies.

Human impact

The controls require meaningful human authorization, recourse, override, limitation disclosure, accessibility review, and impact evaluation for material AI-system changes.

AI assistance disclosure

• AI system used: AltmanAI Model 2.0
• Assistance type: Repository review, governance drafting, issue-form design, release standard design, validation, and pull-request preparation
• Human authorization: Blake Hunter Altman issued ALL CLEAR FOR IMPACT in the controlling conversation on July 13, 2026.
• Human accountability: Final merge and organizational adoption remain attributable to the authenticated AltmanAI repository owner.

P.A.I.H.I.

• Proof: Four scoped files and a reproducible branch comparison.
• Alignment: Direct implementation of ALT-001 and AltmanAI's human-first operating model.
• Integrity: The PR makes no claim that documentation alone guarantees safety or compliance.
• Humanity: Human authority, dignity, privacy, accessibility, recourse, and safety are explicit release requirements.
• Impact: Establishes reusable controls that future AltmanAI repositories and AI systems can adopt.

Decision class

Governance-critical.

Authorization

All Clear for Impact: Approved