Unify Project Governance and AI Governance into single framework

[JohnRDOrazio]

Closes #8

Summary

• Merge AI vetting criteria into general Project Vetting Criteria (v0.2) — AI-specific content preserved as clearly marked "AI domain extension" subsections within each criterion, keeping all Magisterial citations, real-life examples (Father Justin, Robodebt, COMPAS), and full bibliography
• Generalize fragmentation research memo from AI-only to all Catholic digital governance — Liturgy, Scripture, Magisterial documents, Canon Law, canonical structures — with AI preserved as detailed case study
• Reorganize file structure — remove ai-governance/, move research memos to research/, update all build scripts, deploy workflow, README, and cross-references

Test plan

• Verify npm run lint:md passes (0 errors)
• Verify all internal .md cross-references resolve correctly
• Verify build scripts reference the correct file paths (build-standalone-html.sh, build-combined-pdf.sh)
• Verify deploy workflow (deploy-docs.yml) references the correct document paths and titles
• Review unified project-governance/project-vetting-criteria.md for completeness — no citations or examples lost from the former AI vetting criteria
• Review research/fragmented-catholic-digital-governance.md for coherence of the broadened framing

🤖 Generated with Claude Code

[JohnRDOrazio]

Preview: Unified Governance Documentation (PDF)

To make it easier to review the final result of these changes without reading diffs, here is a combined PDF build of the full document stack as it stands on this branch:

📄 cdcf-governance-docs.pdf (71 pages)

What changed in the research memos

Former document	New document	Key change
ai-governance/ai-vetting-criteria.md	Merged into project-governance/project-vetting-criteria.md (v0.2)	AI content preserved as "AI domain extension" subsections
ai-governance/fragmented-catholic-ai-governance.md	research/fragmented-catholic-digital-governance.md	Broadened to cover liturgy, Scripture, Magisterium, canonical structures
ai-governance/governance-as-code-catholic-ai.md	research/governance-as-code-catholic-technology.md	Broadened to all technology deployment + standards compliance
ai-governance/trusted-synthetic-data-ministry-ai.md	research/trusted-data-infrastructure-catholic-ministry.md	Reframed as dual data challenge: institutional (synthetic) + ecclesial (standardization)

[JohnRDOrazio]

@mj3b before we add any further documents, I would like to propose this merger that I had mentioned, please see the attached pdf in the previous comment and add any comments you may have to this PR, if anything needs adjusting or adaptation before we merge and republish.

[JohnRDOrazio]

Three Tiers of Catholic Data

The trusted data infrastructure memo now distinguishes three tiers of Catholic data, each requiring a different infrastructure model:

Tier	Data Category	Examples	Infrastructure Model	CDCF Role
Tier 1	Sensitive institutional data	Patient records, student files, case management, sacramental records of individuals	Federated synthetic data — never pooled	Sets validation standards and certification criteria
Tier 2	Local institutional data	Diocesan operations, staffing, enrollment, facility data	Locally governed — not a commons	Defines interoperability standards
Tier 3	Universal Church data	Liturgy, Scripture, Magisterium, Canon Law, canonical structures	A genuine data commons	Defines standards, maintains authoritative datasets, serves the shared baseline with contributed local adaptations

The key insight: the blanket "federated, not pooled" response is correct for Tiers 1 and 2, but wrong for Tier 3. The General Roman Calendar, the structure of Sacred Scripture, the documents of the Magisterium, and the directory of dioceses belong to the universal Church — they are not any institution's private data. The CDCF can centrally serve this shared baseline, with local adaptations (proper calendars, vernacular editions, diocesan supplements) contributed by the bishops' conferences and dioceses that govern them.

[JohnRDOrazio]

Latest: copyright-aware hybrid model + bylaws grounding

Two new commits:

Hybrid model for copyrighted local adaptations. Tier 3 (universal Church data) now recognizes that vernacular translations and local liturgical adaptations often fall under copyright — protecting both the authority of the local institution and the rights of the publisher it commissioned. Centrally serving this data does not require open-sourcing it. Two options:

1. Federated (default) — the local institution serves its own data locally using CDCF standard schemas
2. Centralized by agreement — the CDCF serves it centrally, protecting copyrighted data according to norms set by the rights holder

Both use identical schemas, so downstream interoperability is unaffected by the hosting choice.

CDCF charter grounding. The CDCF Role section now explicitly references the Foundation's bylaws ("coordinate, develop, steward, and disseminate open-source software, data repositories, technical standards, and digital platforms") and manifesto ("builder commons") to anchor the three-tier data model in the Foundation's charter.

---

Relationship to PR #7

This PR supersedes #7. PR #7 introduced two changes to the trusted synthetic data memo:

• Reframed the CDCF role from "not a data commons" to "federated data commons," grounded in the bylaws and manifesto
• Added a 4th contribution row: "shared data repository stewardship"

Both of those changes are now incorporated here within a more comprehensive framework: the three-tier data model (Tier 1 federated/synthetic, Tier 2 local, Tier 3 genuine commons with hybrid hosting), the copyright-aware hybrid model, and the bylaws/manifesto grounding. PR #7 can be closed once this PR is merged.