Deep Investigative Research

Investigates complex narratives spanning multiple domains by systematically verifying claims, resolving contradictions, and mapping institutional lineages.

When to Use

• Complex narratives spanning multiple domains (tech, theology, policy, finance)
• Claims involving institutional actors and public figures
• Apparent contradictions requiring resolution
• Historical context needed to assess significance
• Hidden stakeholder incentives driving observable behavior
• Need to separate verified fact from speculation

When NOT to Use

• Simple factual lookups
• No primary sources available
• Purely technical/domain-specific questions
• Time constraints prohibit iterative deepening

---

11 Core Subprocesses

1. Claim Inventory & Categorization

Parse the narrative into atomic, verifiable claims and assign categories:

• Institutional claims -- Official releases, filings, legal records
• Attribution claims -- Quotes, statements by named figures
• Relational claims -- Partnerships, allegations of influence
• Speculative framings -- Interpretations, comparative metaphors

Output: Prioritized verification checklist.

Example:
- "May 15, 2026 marks 135 years since Rerum Novarum" -> Historical fact, calculable
- "Chris Olah spoke at Vatican event" -> Requires event documentation
- "Anthropic-xAI partnership" -> Requires corporate filing verification

2. Source Authority Hierarchy

Evaluate evidence weight using a three-tier system:

Tier	Sources	Sufficiency
1 -- Primary/Institutional	Official press releases, encyclical texts, SEC filings, government records, direct video/audio	Standalone sufficient for factual claims
2 -- Credible Secondary	Established news (NPR, FT, Guardian, NYT), industry analysis, academic journals, trade publications	Require corroboration across independent outlets
3 -- Tertiary/Aggregation	Blog posts, Reddit, reposted content, think tank summaries, commentary	Only for identifying leads; must be verified by Tier 1/2

3. Historical Context Alignment

Check if new claims fit within documented institutional patterns:

• Doctrinal continuity -- Does new teaching align with previous doctrine? (e.g., Magnifica Humanitas vs. Rerum Novarum)
• Precedent matching -- Have similar positions been taken before? (e.g., Vatican's prior AI ethics statements)
• Timeline plausibility -- Do dates/sequencing make institutional sense? (e.g., 135-year anniversary alignment)

Resolution metric: Degree of alignment between claimed position and institutional history.

4. Multi-Stakeholder Contradiction Analysis

Identify and reconcile apparent paradoxes:

Type A -- Moral/Financial Paradox:

• Claimed position vs. observed financial incentives
• Resolution: Document both as simultaneously true; analyze incentive structures

Type B -- Infrastructure/Sovereignty Paradox:

• Claimed regulatory position vs. observed infrastructure choices
• Resolution: Identify structural contradictions

Type C -- Attribution/Knowledge Paradox:

• Claimed statements vs. available documentation
• Resolution: Search for primary transcripts and corroborating reports

5. Infrastructure Partnership Verification

Audit corporate/technical relationships across dimensions:

Dimension	Method
Partnership existence	SEC filings, press releases, company blogs
Technical specifics	Compute capacity specs, infrastructure details
Financial scale	Deal announcements, investor presentations
Geopolitical implications	Government/defense integration tracking
Contradiction with stated values	Cross-reference with ethical statements

6. Iterative Deepening Protocol

Five research cycles building from surface to systemic:

Cycle	Question	Focus
1	Does it exist?	Document/event existence, person/institution reality, date accuracy
2	What exactly was claimed?	Exact wording, attendees, positions, primary source substantiation
3	What contradictions exist?	Tensions between claims and actions, incentive structures
4	How does this fit history?	Institutional lineage, precedents, novelty assessment
5	What's the systemic impact?	Geopolitical/financial consequences, ecosystem response, second-order effects

7. Doctrinal/Theological Lineage Mapping

Trace principles across centuries:

Rerum Novarum (1891)          ->  Magnifica Humanitas (2026)
- Factory workers              ->  Content moderators, data miners
- Steam engines               ->  Neural networks
- Wage exploitation           ->  Data extraction, algorithmic control
- Duty to protect labor       ->  Duty to protect human dignity in AI age
- Worker dignity              ->  Protection against reduction to data points

Method: Identify core principle -> trace through intervening decades -> map to new domain -> document vocabulary evolution.

8. Attribution Verification

Confirm direct quotes and public statements:

1. Identify source claim: "X said Y"
2. Search for primary documentation (official statements, transcripts, interviews, video)
3. Examine context: public vs. private, formal vs. spontaneous, direct quote vs. paraphrase
4. Check for amplification: official confirmations, independent corroboration, corrections

9. Hidden Human Cost Documentation

Verify claims about invisible labor supply chains:

• Content moderation -- Geographic distribution, compensation, psychological impact
• Rare earth extraction -- Mining locations, labor allegations, environmental impact
• Data annotation -- Global workforce distribution, working conditions, compensation

Sources: Labor rights organizations, academic studies, investigative journalism, NGO documentation.

10. Geopolitical Friction Point Identification

Map where moral teachings collide with strategic interests:

• Military/Defense -- Encyclical positions vs. defense procurement and contracts
• Financial/Market -- External guidance claims vs. IPO momentum and valuations
• Sovereignty/Regulation -- State-level regulation claims vs. orbital platforms escaping jurisdiction

11. Metaphorical/Theological Framework Analysis

Extract moral logic underlying policy positions:

Babel Model                      Jerusalem Model
|                                |
Centralized efficiency           Localized collaboration
Data homogenization              Human-centered diversity
Optimization as goal             Dignity as goal
Humans as productivity units     Humans as irreplaceable persons
|                                |
Technological dominance          Technological humility

---

Quality Assurance

Cross-verification

• No single claim accepted without corroboration
• Minimum 2 independent sources for factual assertions
• Institutional claims verified against official releases

Contradiction Documentation

• All tensions explicitly listed, not glossed
• Both sides presented fairly
• Incentive structures analyzed to explain gaps

Temporal Grounding

• All dates verified against historical records
• Timeline consistency checked
• Anniversary alignments flagged as intentional/coincidental

Source Dependency Mapping

• For major conclusions, identify enabling sources
• Note if conclusion relies on single source
• Flag areas of genuine ambiguity

---

Output Structure

Each investigation produces:

1. Factual Verification Summary

• What is confirmed in primary sources
• What is speculative
• What remains unverified

2. Contradiction Map

• Identified tensions with resolution approach
• Acknowledged paradoxes
• Unresolved gaps

3. Historical Context

• Institutional lineage
• Precedents
• Novel elements

4. Stakeholder Position Matrix

• What each actor claims
• What incentives they face
• What actions they're taking
• Gaps between claims and actions

5. Geopolitical Impact Assessment

• Where moral teaching meets strategic interests
• Likely flashpoints
• Unacknowledged consequences