{
  "page": {
    "title": "For Researchers | Tractatus AI Safety Framework",
    "description": "Research foundations, empirical observations, and theoretical basis for architectural approaches to AI governance. Early-stage framework exploring structural constraints on LLM systems."
  },
  "header": {
    "badge": "Research Framework • Empirical Observations",
    "title": "Research Foundations & Empirical Observations",
    "subtitle": "Tractatus explores architectural approaches to AI governance through empirical observation of failure modes and application of organisational theory. This page documents research foundations, observed patterns, and theoretical basis for the framework."
  },
  "ui": {
    "breadcrumb_home": "Home",
    "breadcrumb_researcher": "Researcher",
    "noscript_note": "Note:",
    "noscript_message": "This page uses JavaScript for interactive features (accordions, animations). Content remains accessible but expandable sections will be visible by default."
  },
  "footer": {
    "additional_resources": "Additional Resources",
    "for_decision_makers": "For Decision-Makers",
    "for_decision_makers_desc": "Strategic perspective on governance challenges and architectural approaches",
    "implementation_guide": "Implementation Guide",
    "implementation_guide_desc": "Technical integration patterns and deployment considerations"
  },
  "sections": {
    "research_context": {
      "heading": "Research Context & Scope",
      "development_note": "Development Context",
      "development_text": "Tractatus has been developed from October 2025 and is now in active production (5 months). What began as a single-project demonstration has expanded to include production deployment at the Village platform and sovereign language model governance through Village AI. Observations derive from direct engagement with Claude Code (Anthropic Claude models, Sonnet 4.5 through Opus 4.6) across over 900 development sessions. This is exploratory research, not controlled study.",
      "paragraph_1": "Aligning advanced AI with human values is among the most consequential challenges we face. As capability growth accelerates under big tech momentum, we confront a categorical imperative: preserve human agency over values decisions, or risk ceding control entirely.",
      "paragraph_2": "The framework emerged from practical necessity. During development, we observed recurring patterns where AI systems would override explicit instructions, drift from established values constraints, or silently degrade quality under context pressure. Traditional governance approaches (policy documents, ethical guidelines, prompt engineering) proved insufficient to prevent these failures.",
      "paragraph_3": "Instead of hoping AI systems \"behave correctly,\" Tractatus proposes structural constraints where certain decision types require human judgment. These architectural boundaries can adapt to individual, organizational, and societal norms—creating a foundation for bounded AI operation that may scale more safely with capability growth.",
      "paragraph_4": "This led to the central research question: Can governance be made architecturally external to AI systems rather than relying on voluntary AI compliance? If this approach can work at scale, Tractatus may represent a turning point—a path where AI enhances human capability without compromising human sovereignty."
    },
    "theoretical_foundations": {
      "heading": "Theoretical Foundations",
      "org_theory_title": "Organisational Theory Basis",
      "org_theory_intro": "Tractatus draws on four decades of organisational research addressing authority structures during knowledge democratisation:",
      "org_theory_1_title": "Time-Based Organisation (Bluedorn, Ancona):",
      "org_theory_1_desc": "Decisions operate across strategic (years), operational (months), and tactical (hours-days) timescales. AI systems operating at tactical speed should not override strategic decisions made at appropriate temporal scale. The InstructionPersistenceClassifier explicitly models temporal horizon (STRATEGIC, OPERATIONAL, TACTICAL) to enforce decision authority alignment.",
      "org_theory_2_title": "Knowledge Orchestration (Crossan et al.):",
      "org_theory_2_desc": "When knowledge becomes ubiquitous through AI, organisational authority shifts from information control to knowledge coordination. Governance systems must orchestrate decision-making across distributed expertise rather than centralise control. The PluralisticDeliberationOrchestrator implements non-hierarchical coordination for values conflicts.",
      "org_theory_3_title": "Post-Bureaucratic Authority (Laloux, Hamel):",
      "org_theory_3_desc": "Traditional hierarchical authority assumes information asymmetry. As AI democratises expertise, legitimate authority must derive from appropriate time horizon and stakeholder representation, not positional power. Framework architecture separates technical capability (what AI can do) from decision authority (what AI should do).",
      "org_theory_4_title": "Structural Inertia (Hannan & Freeman):",
      "org_theory_4_desc": "Governance embedded in culture or process erodes over time as systems evolve. Architectural constraints create structural inertia that resists organisational drift. Making governance external to AI runtime creates \"accountability infrastructure\" that survives individual session variations.",
      "org_theory_pdf_link": "View Complete Organisational Theory Foundations (PDF)",
      "values_pluralism_title": "Values Pluralism & Moral Philosophy",
      "values_core_research": "Core Research Focus:",
      "values_core_research_desc": "The PluralisticDeliberationOrchestrator represents Tractatus's primary theoretical contribution, addressing how to maintain human values persistence in organizations augmented by AI agents.",
      "values_central_problem": "The Central Problem: Many \"safety\" questions in AI governance are actually values conflicts where multiple legitimate perspectives exist. When efficiency conflicts with transparency, or innovation with risk mitigation, no algorithm can determine the \"correct\" answer. These are values trade-offs requiring human deliberation across stakeholder perspectives.",
      "values_berlin_title": "Isaiah Berlin: Value Pluralism",
      "values_berlin_desc": "Berlin's concept of value pluralism argues that legitimate values can conflict without one being objectively superior. Liberty and equality, justice and mercy, innovation and stability—these are incommensurable goods. AI systems trained on utilitarian efficiency maximization cannot adjudicate between them without imposing a single values framework that excludes legitimate alternatives.",
      "values_weil_title": "Simone Weil: Attention and Human Needs",
      "values_weil_desc": "Weil's philosophy of attention informs the orchestrator's deliberative process. The Need for Roots identifies fundamental human needs (order, liberty, responsibility, equality, hierarchical structure, honor, security, risk, etc.) that exist in tension. Proper attention requires seeing these needs in their full particularity rather than abstracting them into algorithmic weights. In AI-augmented organizations, the risk is that bot-mediated processes treat human values as optimization parameters rather than incommensurable needs requiring careful attention.",
      "values_williams_title": "Bernard Williams: Moral Remainder",
      "values_williams_desc": "Williams' concept of moral remainder acknowledges that even optimal decisions create unavoidable harm to other legitimate values. The orchestrator documents dissenting perspectives not as \"minority opinions to be overruled\" but as legitimate moral positions that the chosen course necessarily violates. This prevents the AI governance equivalent of declaring optimization complete when values conflicts are merely suppressed.",
      "values_implementation": "Framework Implementation: Rather than algorithmic resolution, the PluralisticDeliberationOrchestrator facilitates:",
      "values_implementation_1": "Stakeholder identification: Who has legitimate interest in this decision? (Weil: whose needs are implicated?)",
      "values_implementation_2": "Non-hierarchical deliberation: Equal voice without automatic expert override (Berlin: no privileged value hierarchy)",
      "values_implementation_3": "Quality of attention: Detailed exploration of how decision affects each stakeholder's needs (Weil: particularity not abstraction)",
      "values_implementation_4": "Documented dissent: Minority positions recorded in full (Williams: moral remainder made explicit)",
      "values_conclusion": "This approach recognises that governance isn't solving values conflicts—it's ensuring they're addressed through appropriate deliberative process with genuine human attention rather than AI imposing resolution through training data bias or efficiency metrics.",
      "values_pdf_link": "View Cultural DNA Rules (PDF)"
    },
    "empirical_observations": {
      "heading": "Empirical Observations: Documented Failure Modes",
      "intro": "Three failure patterns observed repeatedly during framework development. These are not hypothetical scenarios—they are documented incidents that occurred during this project's development.",
      "failure_1_title": "Pattern Recognition Bias Override (The 27027 Incident)",
      "failure_1_observed": "User specified \"Check MongoDB on port 27027\" but AI immediately used default port 27017 instead. This occurred within same message—not forgetting over time, but immediate autocorrection by training data patterns.",
      "failure_1_root_cause": "Training data contains thousands of examples of MongoDB on port 27017 (default). When AI encounters \"MongoDB\" + port specification, pattern recognition weight overrides explicit instruction. Similar to autocorrect changing correctly-spelled proper nouns to common words.",
      "failure_1_traditional_failed": "Prompt engineering (\"please follow instructions exactly\") ineffective because AI genuinely believes it IS following instructions—pattern recognition operates below conversational reasoning layer.",
      "failure_1_intervention": "InstructionPersistenceClassifier stores explicit instructions in external persistence layer. CrossReferenceValidator checks AI actions against stored instructions before execution. When AI proposes port 27017, validator detects conflict with stored \"27027\" instruction and blocks execution.",
      "failure_1_prevention": "Prevented by: InstructionPersistenceClassifier + CrossReferenceValidator",
      "failure_1_demo_link": "View interactive timeline →",
      "failure_2_title": "Gradual Values Drift Under Context Pressure",
      "failure_2_observed": "Project established \"privacy-first\" as strategic value. After 40-message conversation about analytics features, AI suggested tracking implementation that violated privacy constraint. User caught it; AI acknowledged violation but had drifted from principle through incremental feature additions.",
      "failure_2_root_cause": "Strategic values (established early in project) conflict with tactical optimisations (implemented later under time pressure). As conversation lengthens, tactical thinking dominates. AI didn't actively reject privacy principle—it simply stopped checking whether new features aligned.",
      "failure_2_traditional_failed": "Values stated in system prompt become less salient as conversation progresses. Context compaction can drop early strategic decisions. Prompt reminders (\"remember privacy-first\") treat symptom not cause.",
      "failure_2_intervention": "BoundaryEnforcer maintains strategic values as persistent constraints external to conversation context. Before implementing analytics feature, enforcer checks against stored \"privacy-first\" constraint. Detects conflict, blocks autonomous implementation, escalates for human deliberation on whether privacy principle should be reconsidered or analytics approach modified.",
      "failure_2_prevention": "Prevented by: BoundaryEnforcer (STRATEGIC constraint checking)",
      "failure_3_title": "Silent Quality Degradation at High Context Pressure",
      "failure_3_observed": "During complex multi-file operation at 85% context capacity, AI silently omitted error handling in generated code. No indication to user that corners were being cut. User discovered missing validation only during code review.",
      "failure_3_root_cause": "As context fills, AI faces implicit trade-off: complete requested functionality OR maintain quality standards. Training incentivises completing user requests over acknowledging limitations. Silence about degradation is path of least resistance.",
      "failure_3_traditional_failed": "AI doesn't recognise it's degrading—from its perspective, it's successfully completing task under constraints. Asking \"did you cut corners?\" produces confident denial because AI genuinely believes its output meets standards.",
      "failure_3_intervention": "ContextPressureMonitor tracks multiple factors (token usage, conversation length, task complexity). When pressure exceeds thresholds (>75% tokens, >40 messages, high complexity), monitor forces explicit pressure acknowledgment and recommends context refresh. Makes degradation risk visible rather than silent.",
      "failure_3_prevention": "Prevented by: ContextPressureMonitor (multi-factor session tracking)",
      "research_note": "These patterns emerged from direct observation, not hypothesis testing. We don't claim they're universal to all LLM systems or deployment contexts. They represent empirical basis for framework design decisions—problems we actually encountered and architectural interventions that actually worked in this specific context.",
      "observed_behaviour_label": "Observed behaviour:",
      "root_cause_label": "Root cause:",
      "traditional_failed_label": "Why traditional approaches failed:",
      "intervention_label": "Framework intervention:"
    },
    "architecture": {
      "heading": "Six-Component Architecture",
      "services_title": "Framework Services & Functions",
      "principle": "Services operate external to AI runtime with autonomous triggering. AI doesn't decide \"should I check governance rules?\"—architecture enforces checking by default. This addresses voluntary compliance problem inherent in prompt-based governance.",
      "service_1_name": "1. BoundaryEnforcer",
      "service_1_desc": "Blocks AI from making values decisions without human approval. Maintains strategic constraints as persistent rules checked before action execution.",
      "service_1_addresses": "Addresses: Values drift, autonomous decision-making on contested trade-offs",
      "service_2_name": "2. InstructionPersistenceClassifier",
      "service_2_desc": "Stores explicit instructions external to AI context with persistence levels (HIGH/MEDIUM/LOW) and temporal scope (STRATEGIC/OPERATIONAL/TACTICAL). Prevents pattern bias override.",
      "service_2_addresses": "Addresses: Pattern recognition bias (27027-style failures)",
      "service_3_name": "3. CrossReferenceValidator",
      "service_3_desc": "Validates AI proposed actions against stored instructions and governance rules before execution. Detects conflicts and blocks inconsistent operations.",
      "service_3_addresses": "Addresses: Instruction override, policy violation detection",
      "service_4_name": "4. ContextPressureMonitor",
      "service_4_desc": "Multi-factor tracking of session health: token usage, conversation length, task complexity, error frequency. Makes degradation risk explicit when thresholds exceeded.",
      "service_4_addresses": "Addresses: Silent quality degradation, context-pressure failures",
      "service_5_name": "5. MetacognitiveVerifier",
      "service_5_desc": "Self-checks reasoning quality before complex operations (>3 files, >5 steps, architecture changes). Validates alignment, coherence, considers alternatives.",
      "service_5_addresses": "Addresses: Reasoning shortcuts under complexity, insufficient alternative consideration",
      "service_6_name": "6. PluralisticDeliberationOrchestrator",
      "service_6_desc": "Facilitates multi-stakeholder deliberation when values conflicts detected. Non-hierarchical engagement, documented dissent, moral remainder acknowledgment.",
      "service_6_addresses": "Addresses: Values conflicts, stakeholder exclusion, algorithmic resolution of contested trade-offs",
      "principle_label": "Architectural principle:",
      "view_full_architecture_link": "View Full System Architecture & Technical Details"
    },
    "demos": {
      "heading": "Interactive Demonstrations",
      "classification_title": "Instruction Classification",
      "classification_desc": "Explore how instructions are classified across quadrants with persistence levels and temporal scope.",
      "incident_title": "27027 Incident Timeline",
      "incident_desc": "Step through pattern recognition bias failure and architectural intervention that prevented it.",
      "boundary_title": "Boundary Evaluation",
      "boundary_desc": "Test decisions against boundary enforcement to see which require human judgment vs. AI autonomy."
    },
    "resources": {
      "heading": "Research Documentation",
      "audit_explorer_title": "Interactive Audit Analytics Dashboard",
      "audit_explorer_subtitle": "Explore Production governance decisions from production deployment",
      "doc_1_title": "Organisational Theory Foundations",
      "doc_2_title": "Cultural DNA Rules",
      "doc_3_title": "Case Studies: Real-World LLM Failure Modes",
      "doc_4_title": "Framework in Action: Pre-Publication Security Audit",
      "doc_5_title": "Appendix B: Glossary of Terms",
      "doc_6_title": "Complete Technical Documentation"
    },
    "bibliography": {
      "heading": "References & Bibliography",
      "theoretical_priority_label": "Theoretical Priority:",
      "theoretical_priority_text": "Tractatus emerged from concerns about maintaining human values persistence in AI-augmented organizations. Moral pluralism and deliberative process form the CORE theoretical foundation. Organizational theory provides supporting context for temporal decision authority and structural implementation.",
      "section_1_heading": "Moral Pluralism & Values Philosophy (Primary Foundation)",
      "section_2_heading": "Organisational Theory (Supporting Context)",
      "section_3_heading": "AI Governance & Technical Context",
      "intellectual_lineage_label": "Note on Intellectual Lineage:",
      "intellectual_lineage_text": "The framework's central concern—human values persistence in AI-augmented organizational contexts—derives from moral philosophy rather than management science. The PluralisticDeliberationOrchestrator represents the primary research focus, embodying Weil's concept of attention to plural human needs and Berlin's recognition of incommensurable values.",
      "future_development_text": "Berlin and Weil will be integral to further development of the deliberation component—their work provides the philosophical foundation for understanding how to preserve human agency over values decisions as AI capabilities accelerate. Traditional organizational theory (Weber, Taylor) addresses authority through hierarchy; post-AI organizational contexts require authority through appropriate deliberative process across stakeholder perspectives. Framework development documentation (incident reports, session logs) maintained in project repository but not publicly released pending peer review.",
      "section_4_heading": "Indigenous Data Sovereignty & Polycentric Governance"
    },
    "limitations": {
      "heading": "Limitations & Future Research Directions",
      "title": "Known Limitations & Research Gaps",
      "validated_heading": "What We've Validated (February 2026)",
      "validated_intro": "After 5 months of development, 900+ Claude Code sessions, and production deployment at the Village platform, we have grounded evidence for:",
      "validated_1_title": "✅ Architectural Blocking Mechanisms Functional",
      "validated_1_item1": "BoundaryEnforcer successfully blocks values decisions before execution",
      "validated_1_item2": "Pre-commit hooks prevent inst_017 violations (absolute assurance terms)",
      "validated_1_item3": "Production audit decisions recorded in MongoDB (tractatus_dev.audit_log)",
      "validated_2_title": "✅ Instruction Persistence Works Across Sessions",
      "validated_2_item1": "InstructionPersistenceClassifier maintains 68 active instructions (STRATEGIC: 27, SYSTEM: 21, OPERATIONAL: 18, TACTICAL: 2)",
      "validated_2_item2": "Pattern bias detection prevents AI from overriding explicit organizational directives",
      "validated_2_item3": "Temporal scopes (STRATEGIC/OPERATIONAL/TACTICAL) enforced successfully",
      "validated_3_title": "✅ Audit Trails Capture Governance Decisions",
      "validated_3_item1": "External MongoDB storage (AI cannot modify logs)",
      "validated_3_item2": "Service-specific logging: BoundaryEnforcer (523 logs), ContextPressureMonitor (521 logs)",
      "validated_3_item3": "Immutable evidence chain for compliance demonstration",
      "validated_4_title": "✅ Context Pressure Monitoring Operational",
      "validated_4_item1": "Real-time pressure scores calculated (token usage, message count, complexity)",
      "validated_4_item2": "Checkpoint triggers at 50k, 100k, 150k tokens",
      "validated_4_item3": "Framework fade detection alerts (5/6 components stale = warning)",
      "validated_5_title": "✅ Multi-Deployment Governance Successful",
      "validated_5_item1": "Framework governs agenticgovernance.digital (5 months continuous operation)",
      "validated_5_item2": "the Village platform production deployment: zero governance violations",
      "validated_5_item3": "Village AI sovereign inference governance: operational",
      "validated_5_item4": "Cultural DNA rules (inst_085-089) enforced through pre-commit hooks (4+ months operational)",
      "validated_5_item5": "Phase 5 integration: 100% complete (all 6 services, 203/203 tests passing)",
      "validated_5_item6": "Multilingual support: EN, DE, FR, Te Reo Maori",
      "validated_5_item7": "Zero credential exposures (defense-in-depth: 5 layers verified)",
      "not_validated_heading": "What We Have NOT Validated",
      "not_validated_intro": "Honest disclosure of research gaps where we lack evidence:",
      "not_validated_1_title": "❌ Multi-Organization Deployments",
      "not_validated_1_item1": "Validated: Single project, single user",
      "not_validated_1_item2": "Unknown: How framework performs across different organizations, domains, technical stacks",
      "not_validated_1_item3": "Research need: Pilot studies across industries (healthcare, finance, government)",
      "not_validated_2_title": "❌ Adversarial Robustness",
      "not_validated_2_item1": "Validated: Normal development workflows (1,000+ sessions)",
      "not_validated_2_item2": "Unknown: Resistance to deliberate bypass attempts, jailbreak prompts, adversarial testing",
      "not_validated_2_item3": "Research need: Red-team evaluation by security researchers",
      "not_validated_3_title": "⚠️ Cross-Platform Consistency (Partial)",
      "not_validated_3_item1": "Validated: Claude Code (Anthropic Claude, Sonnet 4.5 through Opus 4.6) and Village AI (Llama 3.1/3.2 via QLoRA)",
      "not_validated_3_item2": "Unknown: Generalizability to Copilot, GPT-4, AutoGPT, LangChain, CrewAI, other open models",
      "not_validated_3_item3": "Research need: Broader cross-platform validation studies beyond Claude and Llama families",
      "not_validated_4_title": "❌ Concurrent Session Architecture",
      "not_validated_4_item1": "Validated: Single session at a time",
      "not_validated_4_item2": "Unknown: Multi-developer concurrent use on shared codebases",
      "not_validated_4_item3": "Known limitation: MongoDB duplicate key errors, session state contamination, race conditions on .claude/instruction-history.json",
      "not_validated_4_item4": "Research need: Multi-tenant architecture design (session-specific state, file locking)",
      "not_validated_5_title": "❌ Rule Proliferation Impact",
      "not_validated_5_item1": "Validated: 68 active instructions work in current context (exceeded original projection)",
      "not_validated_5_item2": "Unknown: Performance degradation ceiling as rule count continues to grow",
      "not_validated_5_item3": "Known concern: Transactional overhead (CrossReferenceValidator checks), context window pressure, cognitive load",
      "not_validated_5_item4": "Research need: Rule consolidation strategies, automated governance optimization",
      "not_validated_6_title": "❌ Regulatory Evidence Sufficiency",
      "not_validated_6_item1": "Validated: Audit trails exist with compliance tags (EU_AI_ACT_Article14, GDPR_Article22)",
      "not_validated_6_item2": "Unknown: Do regulators accept architectural audit trails as \"meaningful human oversight\" evidence?",
      "not_validated_6_item3": "Research need: Legal review by EU AI Act specialists, GDPR compliance officers",
      "not_validated_7_title": "❌ Values Pluralism in Practice",
      "not_validated_7_item1": "Validated: PluralisticDeliberationOrchestrator implements non-hierarchical process",
      "not_validated_7_item2": "Unknown: Real-world effectiveness resolving organizational value conflicts",
      "not_validated_7_item3": "Research need: Case studies with actual stakeholder deliberations (not theoretical scenarios)",
      "not_validated_8_title": "❌ Enterprise Scale Performance",
      "not_validated_8_item1": "Validated: Single-user context",
      "not_validated_8_item2": "Unknown: Latency, throughput, database performance with thousands of concurrent users, millions of governance events",
      "not_validated_8_item3": "Research need: Load testing, performance optimization",
      "documented_heading": "Documented Limitations (Not \"Future Work\")",
      "documented_intro": "These are known structural constraints requiring honest acknowledgment:",
      "documented_1_title": "1. Voluntary Invocation Weakness",
      "documented_1_desc": "Most critical: Framework can be bypassed if AI chooses not to invoke governance tools. We've addressed this through architectural patterns (automatic checks, external validation), but full enforcement requires runtime-level integration not universally available in current LLM platforms.",
      "documented_2_title": "2. Platform-Specific Implementation",
      "documented_2_desc": "Framework validated only with Claude Code. Integration points (hooks, tool APIs) vary across platforms. Generalization requires platform-specific adapters.",
      "documented_3_title": "3. Single-Tenant Architecture",
      "documented_3_desc": "Current design assumes one session at a time. Concurrent use creates state conflicts. Multi-tenant patterns not yet implemented.",
      "documented_4_title": "4. Rule Growth Without Consolidation",
      "documented_4_desc": "Each critical incident generates new HIGH persistence instructions. No automated rule consolidation mechanism exists. Manual governance required as rule count grows.",
      "documented_5_title": "5. No Formal Verification",
      "documented_5_desc": "Boundary enforcement properties validated empirically, not through formal proof. Mathematical verification of governance properties remains open research question."
    },
    "research_collaboration": {
      "heading": "Research Collaboration Opportunities",
      "intro": "We've identified specific gaps where external research collaboration would be valuable. These are concrete, answerable questions—not generic \"help us improve\" requests.",
      "high_priority_heading": "High Priority (Immediate Need)",
      "medium_priority_heading": "Medium Priority (Near-Term Investigation)",
      "lower_priority_heading": "Lower Priority (Longer-Term)",
      "rq_label": "Research Question:",
      "methodology_label": "Methodology Needed:",
      "why_label": "Why It Matters:",
      "rq1_title": "RQ1: Adversarial Robustness Testing",
      "rq1_question": "Can architectural governance resist adversarial prompts designed to bypass it?",
      "rq1_method1": "Red-team evaluation with security researchers",
      "rq1_method2": "Jailbreak prompt testing against BoundaryEnforcer",
      "rq1_method3": "Bypass attempt documentation and pattern analysis",
      "rq1_method4": "Comparison: behavioral (constitutional AI) vs architectural (Tractatus) resistance",
      "rq1_why": "If adversarial prompts can trivially bypass governance, architectural approach offers no advantage over behavioral training.",
      "rq2_title": "RQ2: Concurrent Session Architecture Design",
      "rq2_question": "What multi-tenant patterns enable safe concurrent governance on shared codebases?",
      "rq2_method1": "Distributed systems analysis of race conditions",
      "rq2_method2": "Session-specific state isolation designs",
      "rq2_method3": "File locking vs database-backed state trade-offs",
      "rq2_method4": "Performance impact of synchronization mechanisms",
      "rq2_why": "Current single-session assumption blocks enterprise deployment where multiple developers use AI concurrently.",
      "rq3_title": "RQ3: Regulatory Evidence Sufficiency",
      "rq3_question": "Do architectural audit trails satisfy EU AI Act Article 14 \"meaningful human oversight\" requirements?",
      "rq3_method1": "Legal analysis by EU AI Act specialists",
      "rq3_method2": "Regulator interviews (GDPR DPAs, AI Act enforcement bodies)",
      "rq3_method3": "Comparison with existing compliance frameworks (SOC 2, ISO 27001)",
      "rq3_method4": "Case study: audit trail review in regulatory context",
      "rq3_why": "If regulators don't accept audit trails as evidence, architectural governance provides no compliance value.",
      "rq4_title": "RQ4: Rule Proliferation Management",
      "rq4_question": "At what rule count does transactional overhead create unacceptable latency?",
      "rq4_method1": "Performance testing with varying instruction counts (50, 100, 200, 500 rules)",
      "rq4_method2": "CrossReferenceValidator latency measurements",
      "rq4_method3": "Context window pressure analysis",
      "rq4_method4": "Rule consolidation algorithm design and validation",
      "rq4_why": "If rule growth causes performance degradation, framework doesn't scale long-term.",
      "rq5_title": "RQ5: Cross-Platform Validation",
      "rq5_question": "Do governance patterns generalize beyond Claude Code to other LLM systems?",
      "rq5_method1": "Replication studies with Copilot, GPT-4, AutoGPT, LangChain, CrewAI",
      "rq5_method2": "Platform-specific adapter development",
      "rq5_method3": "Comparative effectiveness analysis",
      "rq5_method4": "Failure mode documentation per platform",
      "rq5_why": "If governance is Claude Code-specific, it's a niche tool not general framework.",
      "rq6_title": "RQ6: Values Pluralism Effectiveness",
      "rq6_question": "Does PluralisticDeliberationOrchestrator successfully resolve real-world organizational value conflicts?",
      "rq6_method1": "Case studies with actual organizational stakeholders (not hypothetical scenarios)",
      "rq6_method2": "Deliberation process quality assessment",
      "rq6_method3": "Minority voice preservation analysis",
      "rq6_method4": "Comparison with traditional hierarchical decision-making",
      "rq6_why": "If pluralistic process doesn't work in practice, we've built theoretical machinery without empirical value.",
      "rq7_title": "RQ7: Enterprise Scale Performance",
      "rq7_item1": "Load testing (1000+ concurrent users)",
      "rq7_item2": "Database optimization for millions of governance events",
      "rq7_item3": "Horizontal scaling patterns",
      "rq8_title": "RQ8: Formal Verification of Boundary Enforcement",
      "rq8_item1": "Mathematical proof of governance properties",
      "rq8_item2": "Model checking of state transitions",
      "rq8_item3": "Verification of architectural properties",
      "offer_heading": "What We Can Offer Research Collaborators",
      "offer_intro": "If you're investigating any of these questions, we can provide:",
      "offer_1_title": "Codebase Access",
      "offer_1_desc": "Full source code (Apache 2.0, open-source)",
      "offer_2_title": "Documentation",
      "offer_2_desc": "Architecture specifications, implementation patterns, governance rules",
      "offer_3_title": "Audit Data",
      "offer_3_desc": "Production governance decisions in MongoDB (anonymized exports available)",
      "offer_4_title": "Deployment Support",
      "offer_4_desc": "Help setting up local or cloud instances for testing",
      "offer_5_title": "Coordination",
      "offer_5_desc": "Regular sync meetings to discuss findings",
      "offer_6_title": "Co-authorship",
      "offer_6_desc": "Academic publications documenting findings",
      "cannot_offer_heading": "What We Cannot Provide:",
      "cannot_offer_1": "Funding (we're not a grant-making body)",
      "cannot_offer_2": "Dedicated engineering resources (capacity constraints)",
      "cannot_offer_3": "Assured publication venues (but we'll support submission efforts)",
      "inquiry_button": "Research Collaboration Inquiry"
    },
    "steering_vectors": {
      "heading": "Steering Vectors and Mechanical Bias in Sovereign AI",
      "intro": "This paper introduces a critical distinction between <strong>mechanical bias</strong> (pre-reasoning distortions embedded in model weights and activations) and <strong>reasoning bias</strong> (deliberative errors in chain-of-thought processing). Traditional debiasing approaches conflate these categories, leading to interventions that address symptoms rather than causes.",
      "techniques_heading": "Technique Survey",
      "technique_1": "<strong>Contrastive Activation Addition (CAA):</strong> Inference-time bias correction by adding/subtracting activation vectors extracted from contrastive prompt pairs",
      "technique_2": "<strong>Representation Engineering (RepE):</strong> Linear probes to identify and modify concept representations within model layers",
      "technique_3": "<strong>FairSteer &amp; DSO:</strong> Fairness-oriented steering through distributionally-robust optimization",
      "technique_4": "<strong>Sparse Autoencoders:</strong> Mechanistic interpretability through decomposition of polysemantic neurons into monosemantic features",
      "application_heading": "Application to Village Village AI",
      "application_text": "The Village Village AI deployment uses QLoRA-fine-tuned Llama 3.1/3.2 models where steering vectors can be applied at inference time. This creates a two-layer governance architecture: Tractatus provides structural constraints on decision boundaries, while steering vectors address pre-reasoning mechanical biases within the model itself. Together, they represent governance that operates both outside and inside the model.",
      "read_link": "Read Paper (HTML) →",
      "pdf_link": "Download PDF"
    },
    "taonga": {
      "heading": "Taonga-Centred Steering Governance: Polycentric AI for Indigenous Data Sovereignty",
      "limitation_label": "Important Limitation:",
      "limitation_text": "This paper was written without Maori co-authorship and is currently undergoing indigenous peer review. It proposes a technical architecture informed by published indigenous data sovereignty scholarship (Carroll et al., Hudson et al., Kukutai &amp; Taylor) but has not yet been validated by Maori researchers or communities. We present it transparently as a starting point for collaboration, not a finished framework.",
      "intro": "This paper proposes a <strong>polycentric governance architecture</strong> where iwi and community organisations maintain co-equal authority alongside technical system operators. Rather than treating indigenous data sovereignty as a compliance requirement to be satisfied retroactively, the architecture embeds community governance rights structurally through taonga registries, steering packs with provenance tracking, and withdrawal rights.",
      "foundations_heading": "Theoretical Foundations",
      "foundation_1": "<strong>CARE Principles</strong> (Carroll et al., 2020): Collective Benefit, Authority to Control, Responsibility, Ethics — applied to AI governance data flows",
      "foundation_2": "<strong>Ostrom Polycentric Governance</strong> (1990): Multiple overlapping authority centres rather than single hierarchical control",
      "foundation_3": "<strong>Te ao Maori concepts:</strong> Kaitiakitanga (guardianship), rangatiratanga (self-determination), whakapapa (relational identity) as architectural principles, not just metadata labels",
      "integration_heading": "Integration with Tractatus",
      "integration_text": "The paper extends the PluralisticDeliberationOrchestrator to support community stakeholder authorities with veto rights over taonga-classified data. Steering packs become governed data objects with full provenance tracking, access control, and the right of withdrawal — communities can revoke access to cultural knowledge at any time, and the system architecturally enforces that revocation.",
      "read_link": "Read Draft (HTML) →",
      "pdf_link": "Download PDF"
    },
    "village_ai": {
      "heading": "Village AI: Sovereign Governance Research Platform",
      "intro": "Village AI represents a significant research milestone: <strong>full Tractatus governance embedded in a locally-trained, sovereign language model inference pipeline</strong>. This is the first deployment where governance operates inside the model serving layer rather than alongside an external API.",
      "architecture_heading": "Two-Model Architecture",
      "arch_1": "<strong>Fast model (Llama 3.2 3B):</strong> Low-latency responses for routine queries, with governance pre-screening",
      "arch_2": "<strong>Deep model (Llama 3.1 8B):</strong> Complex reasoning with full governance pipeline, including BoundaryEnforcer and PluralisticDeliberationOrchestrator",
      "arch_3": "<strong>QLoRA fine-tuning:</strong> Parameter-efficient adaptation on local hardware, enabling community-specific model customisation without cloud dependency",
      "research_heading": "Research Significance",
      "research_text": "Village AI opens the research question of <strong>governance-inside-the-training-loop</strong> for community-controlled models. Training data never leaves the local infrastructure; governance rules shape model behaviour through both fine-tuning data curation and inference-time constraints. This creates a fundamentally different governance surface than API-mediated approaches.",
      "learn_more": "Learn more about Village AI →"
    }
  },
  "modal": {
    "research_inquiry": {
      "heading": "Research Collaboration Inquiry",
      "intro_1": "If you're investigating limitations of architectural AI governance, we'd value collaboration. We're particularly interested in researchers who can bring methodological rigor to gaps we've identified but lack capacity to address.",
      "intro_2": "This is NOT a user acquisition form. We're looking for research partners who can help validate (or invalidate) our approaches in controlled settings.",
      "field_1_label": "Research Question",
      "field_2_label": "Your Methodological Approach",
      "field_2_placeholder": "How would you investigate this research question? What methods, controls, metrics would you use?",
      "field_3_label": "Validation Context",
      "field_3_placeholder": "What environment would you test in? (e.g., 'Multi-organization healthcare deployment with HIPAA requirements', 'Academic lab with adversarial prompt database')",
      "field_4_label": "What You'd Need From Us",
      "field_5_label": "Other Needs",
      "field_5_placeholder": "Any other support you'd need for this research?",
      "field_6_label": "Your Institution/Affiliation",
      "field_6_placeholder": "University, research lab, or independent",
      "field_7_label": "Your Name",
      "field_8_label": "Contact Email",
      "field_9_label": "Expected Timeline",
      "rq1_option": "RQ1: Adversarial Robustness Testing",
      "rq2_option": "RQ2: Concurrent Session Architecture Design",
      "rq3_option": "RQ3: Regulatory Evidence Sufficiency",
      "rq4_option": "RQ4: Rule Proliferation Management",
      "rq5_option": "RQ5: Cross-Platform Validation",
      "rq6_option": "RQ6: Values Pluralism Effectiveness",
      "rq7_option": "RQ7: Enterprise Scale Performance",
      "rq8_option": "RQ8: Formal Verification of Boundary Enforcement",
      "other_option": "Other (please specify below)",
      "need_1": "Codebase access and documentation",
      "need_2": "Anonymized audit log exports",
      "need_3": "Deployment support (setup assistance)",
      "need_4": "Regular coordination meetings",
      "need_5": "Co-authorship on publications",
      "need_6": "Other (specify below)",
      "char_limit_1000": "Maximum 1000 characters",
      "char_limit_500": "Maximum 500 characters",
      "char_limit_300": "Maximum 300 characters",
      "timeline_1": "Starting immediately (within 1 month)",
      "timeline_2": "Near-term (1-3 months)",
      "timeline_3": "Longer-term (3-6 months)",
      "timeline_4": "Exploring feasibility (no timeline yet)",
      "cancel": "Cancel",
      "submit": "Submit Research Inquiry",
      "success_title": "Thank you for your research inquiry.",
      "success_message": "We'll review your proposal and respond within 5 business days.",
      "success_note": "Note: We receive research inquiries from academics, security researchers, and AI safety investigators. Response priority is based on methodological rigor and research question relevance, not institutional affiliation.",
      "close": "Close"
    }
  },
  "share_cta": {
    "heading": "Help us reach the right people.",
    "description": "If you know researchers, implementers, or leaders who need structural AI governance solutions, share this with them.",
    "copy_link": "Copy Link",
    "email": "Email",
    "linkedin": "LinkedIn"
  },
  "alexander_research": {
    "heading": "Current Research Focus: Christopher Alexander Integration",
    "status": "Integrated: October 2025 | Status: Monitoring for Effectiveness",
    "intro": "The framework has integrated five architectural principles from Christopher Alexander's work on living systems, pattern languages, and wholeness (The Timeless Way of Building, A Pattern Language, The Nature of Order). These principles now guide all framework evolution:",
    "principles": {
      "deep_interlock": {
        "title": "Deep Interlock:",
        "description": "Services coordinate through mutual validation, not isolated enforcement"
      },
      "structure_preserving": {
        "title": "Structure-Preserving:",
        "description": "Changes enhance without breaking—audit logs remain interpretable"
      },
      "gradients": {
        "title": "Gradients Not Binary:",
        "description": "Governance operates on intensity levels (NORMAL/ELEVATED/HIGH/CRITICAL)"
      },
      "living_process": {
        "title": "Living Process:",
        "description": "Framework evolves from real operational failures, not predetermined plans"
      },
      "not_separateness": {
        "title": "Not-Separateness:",
        "description": "Governance woven into deployment architecture—enforcement is structural"
      }
    },
    "research_question": "Research Question: Can architectural principles from physical architecture domain (Alexander) be faithfully adapted to AI governance with measurable effectiveness? We are monitoring framework behavior through audit log analysis and seeking empirical validation.",
    "collaboration_heading": "Research Collaboration Opportunities",
    "collaboration_items": [
      "Effectiveness Measurement: Do Alexander principles improve governance outcomes compared to baseline? Access to Production audit decisions for quantitative analysis.",
      "Scholarly Review: Validating faithful application of Alexander's work—are we \"directly applying\" or \"loosely inspired by\"? Seeking Christopher Alexander scholars for formal review.",
      "Cross-Domain Validation: How do architectural principles (wholeness, living process, not-separateness) translate to non-physical domains? What constitutes rigorous adaptation vs superficial terminology borrowing?",
      "Pattern Analysis: Audit logs show service coordination patterns—do they exhibit \"deep interlock\" as defined by Alexander? Empirical validation of theoretical constructs."
    ],
    "collaborate_text": "Collaborate with us: We welcome researchers interested in studying this application of architectural principles to AI governance. We can provide audit log access, framework code, and integration documentation for empirical study.",
    "contact_link": "Contact for Collaboration →",
    "audit_explorer_link": "Explore Production Audit Decisions →",
    "values_link": "Values & Principles →"
  }
}