df8c6ccb03
Replace "ensures", "guarantee", "foolproof", "world-class" and similar absolute terms with evidence-based language throughout public pages, JS components, and FAQ content. Changes apply inst_017 (no absolute assurance terms) consistently. Replacements: - "ensures X" → "validates X", "so that X", "supports X", "maintains X" - "guarantee" → removed or rephrased with qualified language - "foolproof" → "infallible" - "architecturally impossible" → "architecture prevents without explicit override flags" Preserved: published research papers (architectural-alignment*.html), EU AI Act quotes, Te Tiriti treaty language, and FAQ meta-commentary that deliberately critiques this language (lines 2842-2896). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
303 lines
20 KiB
HTML
303 lines
20 KiB
HTML
<!DOCTYPE html>
|
|
<html lang="en" data-page="home-ai">
|
|
<head>
|
|
<meta charset="UTF-8">
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
|
<title>Home AI | Tractatus AI Safety Framework</title>
|
|
<meta name="description" content="Home AI: a sovereign, locally-trained language model operating under Tractatus governance. Six architectural services applied to every interaction.">
|
|
|
|
<meta property="og:type" content="website">
|
|
<meta property="og:url" content="https://agenticgovernance.digital/home-ai.html">
|
|
<meta property="og:title" content="Home AI — Sovereign Language Model | Tractatus">
|
|
<meta property="og:description" content="How a locally-trained language model operates under architectural governance constraints in a production community platform.">
|
|
<meta property="og:image" content="https://agenticgovernance.digital/images/social-preview.png">
|
|
<meta property="og:site_name" content="Tractatus Framework">
|
|
|
|
<meta name="twitter:card" content="summary_large_image">
|
|
<meta name="twitter:title" content="Home AI — Sovereign Language Model | Tractatus">
|
|
<meta name="twitter:description" content="How a locally-trained language model operates under architectural governance constraints in a production community platform.">
|
|
<meta name="twitter:image" content="https://agenticgovernance.digital/images/social-preview.png">
|
|
|
|
<link rel="icon" type="image/svg+xml" href="/favicon-new.svg">
|
|
<link rel="stylesheet" href="/css/fonts.css">
|
|
<link rel="stylesheet" href="/css/tailwind.css">
|
|
<link rel="stylesheet" href="/css/tractatus-theme.min.css">
|
|
<style>
|
|
.skip-link { position: absolute; left: -9999px; top: 0; }
|
|
.skip-link:focus { left: 0; z-index: 100; background: white; padding: 1rem; border: 2px solid #3b82f6; }
|
|
a:focus, button:focus { outline: 3px solid #3b82f6; outline-offset: 2px; }
|
|
a:focus:not(:focus-visible) { outline: none; }
|
|
a:focus-visible { outline: 3px solid #3b82f6; outline-offset: 2px; }
|
|
</style>
|
|
</head>
|
|
<body class="bg-gray-50">
|
|
|
|
<a href="#main-content" class="skip-link">Skip to main content</a>
|
|
|
|
<!-- Navigation -->
|
|
<div id="navbar-placeholder" class="min-h-16"></div>
|
|
<script src="/js/components/navbar.js"></script>
|
|
|
|
<!-- Breadcrumb -->
|
|
<nav class="bg-gray-50 border-b border-gray-200 py-3" aria-label="Breadcrumb">
|
|
<div class="max-w-7xl mx-auto px-4 sm:px-6 lg:px-8">
|
|
<ol class="flex items-center space-x-2 text-sm">
|
|
<li><a href="/" class="text-blue-600 hover:underline transition-colors">Home</a></li>
|
|
<li class="text-gray-400">/</li>
|
|
<li><a href="/village-case-study.html" class="text-blue-600 hover:underline transition-colors">Village Case Study</a></li>
|
|
<li class="text-gray-400">/</li>
|
|
<li class="text-gray-900 font-medium" aria-current="page">Home AI</li>
|
|
</ol>
|
|
</div>
|
|
</nav>
|
|
|
|
<!-- Hero -->
|
|
<header role="banner">
|
|
<section class="bg-gradient-to-br from-teal-700 via-teal-800 to-emerald-800 text-white py-16">
|
|
<div class="max-w-4xl mx-auto px-4 sm:px-6 lg:px-8 text-center">
|
|
<h1 class="text-4xl md:text-5xl font-bold mb-4">Home AI</h1>
|
|
<p class="text-xl text-teal-100 max-w-3xl mx-auto mb-6">
|
|
A sovereign, locally-trained language model operating under architectural governance. Every interaction passes through all six Tractatus services before reaching the user.
|
|
</p>
|
|
<div class="bg-amber-100 border-2 border-amber-400 rounded-lg p-4 max-w-2xl mx-auto">
|
|
<p class="text-amber-900 text-sm">
|
|
<strong>Caveat:</strong> Home AI is a single implementation within the Village platform. Claims below are based on operational observation, not controlled experiments. Independent validation is planned.
|
|
</p>
|
|
</div>
|
|
</div>
|
|
</section>
|
|
</header>
|
|
|
|
<main id="main-content" class="max-w-4xl mx-auto px-4 sm:px-6 lg:px-8 py-16">
|
|
|
|
<!-- What is Home AI -->
|
|
<section class="mb-16">
|
|
<h2 class="text-3xl font-bold text-gray-900 mb-6">What is Home AI?</h2>
|
|
<div class="prose prose-lg text-gray-700">
|
|
<p class="mb-4">
|
|
Home AI is the practical implementation of Tractatus governance within the Village platform — a community-owned digital space where members share stories, documents, and family histories. Unlike cloud-hosted AI assistants, Home AI operates under the principle of <em>digital sovereignty</em>: the community's data and the AI's behaviour are governed by the community itself, not by a remote provider.
|
|
</p>
|
|
<p class="mb-4">
|
|
The term "SLL" (Sovereign Locally-trained Language Model) describes the architectural goal: a language model whose training data, inference, and governance all remain under local control. In practice, Home AI currently uses a hybrid approach — local Llama models for English-language operations and Claude Haiku via API for non-English languages — with a roadmap toward fully local inference as hardware and model capabilities allow.
|
|
</p>
|
|
<p>
|
|
What distinguishes Home AI from other AI assistants is not the model itself, but the governance layer around it. Every interaction — whether a help query, document OCR, story suggestion, or AI-generated summary — passes through the full Tractatus governance stack before any response reaches the user.
|
|
</p>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- The Governance Stack -->
|
|
<section class="mb-16">
|
|
<h2 class="text-3xl font-bold text-gray-900 mb-6">The Governance Stack</h2>
|
|
<p class="text-gray-700 mb-8">
|
|
Each Home AI interaction traverses six governance services in sequence. This is not optional middleware — it operates in the critical execution path, meaning a response cannot be generated without passing through all checks.
|
|
</p>
|
|
|
|
<div class="space-y-4">
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-emerald-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">1. BoundaryEnforcer</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
Detects whether a user query involves values decisions (privacy trade-offs, ethical questions, cultural sensitivity) and blocks the AI from responding autonomously. These are deferred to human moderators. The boundary between "technical question" and "values question" is defined by community-specific rules, not by the AI's judgment.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-indigo-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">2. CrossReferenceValidator</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
Validates the query against stored instructions and known patterns. This is the service that would have caught the 27027 incident — the user's explicit instruction ("use port 27027") is stored externally and cross-referenced against the AI's proposed action ("use port 27017"). When stored instructions conflict with the AI's response, the stored instruction takes precedence.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-amber-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">3. ContextPressureMonitor</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
Tracks session health metrics: token usage, message count, error rate, task complexity. When pressure exceeds thresholds (ELEVATED at 25%, HIGH at 50%, CRITICAL at 75%), the system adjusts validation intensity or recommends session handoff. This prevents the degradation patterns observed in extended AI sessions where error rates compound.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-pink-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">4. MetacognitiveVerifier</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
For complex operations (multi-step tasks, file modifications, configuration changes), the AI performs a structured self-assessment before proposing actions: alignment with instructions, coherence of approach, completeness of plan, safety of proposed changes, and consideration of alternatives. This is triggered selectively to avoid overhead on simple queries.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-blue-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">5. InstructionPersistenceClassifier</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
Classifies instructions by their intended lifespan (HIGH: strategic/permanent, MEDIUM: operational/session-scoped, LOW: tactical/single-use) and quadrant (Strategic, Operational, Tactical, System, Stochastic). This classification determines how strongly the CrossReferenceValidator enforces each instruction and how long it persists in the external store.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-lg shadow-sm p-6 border-l-4 border-teal-500">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-2">6. PluralisticDeliberationOrchestrator</h3>
|
|
<p class="text-gray-700 text-sm">
|
|
When the AI encounters decisions where legitimate values conflict — for example, a member's privacy interests versus community safety concerns — this service halts autonomous decision-making and coordinates a deliberation process among affected stakeholders. The AI presents the conflict and facilitates discussion; it does not resolve it.
|
|
</p>
|
|
</div>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- Governed Features -->
|
|
<section class="mb-16">
|
|
<h2 class="text-3xl font-bold text-gray-900 mb-6">Governed Features</h2>
|
|
<p class="text-gray-700 mb-8">
|
|
Home AI currently provides four AI-powered features, each operating under the full governance stack.
|
|
</p>
|
|
|
|
<div class="grid grid-cols-1 md:grid-cols-2 gap-6">
|
|
<div class="bg-white rounded-xl shadow-md p-6 border border-gray-200">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-3">RAG-Based Help</h3>
|
|
<p class="text-gray-700 text-sm mb-3">
|
|
Vector search retrieves relevant documentation and help content, filtered by the member's permission level. The AI generates contextual answers grounded in retrieved documents rather than from its training data alone.
|
|
</p>
|
|
<p class="text-gray-500 text-xs italic">
|
|
Governance: BoundaryEnforcer prevents PII exposure; CrossReferenceValidator validates responses against platform policies.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-xl shadow-md p-6 border border-gray-200">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-3">Document OCR</h3>
|
|
<p class="text-gray-700 text-sm mb-3">
|
|
Automated text extraction from uploaded documents (historical records, handwritten letters, photographs with text). Extracted text is stored within the member's scope, not shared across tenants or used for model training.
|
|
</p>
|
|
<p class="text-gray-500 text-xs italic">
|
|
Governance: Processing only occurs under explicit consent controls; results are tenant-isolated.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-xl shadow-md p-6 border border-gray-200">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-3">Story Assistance</h3>
|
|
<p class="text-gray-700 text-sm mb-3">
|
|
AI-generated suggestions for writing family stories: prompts, structural advice, and narrative enhancement. Suggestions are filtered through BoundaryEnforcer so that the AI does not impose cultural interpretations or values judgments on family narratives.
|
|
</p>
|
|
<p class="text-gray-500 text-xs italic">
|
|
Governance: Cultural context decisions are deferred to the storyteller, not resolved by the AI.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-white rounded-xl shadow-md p-6 border border-gray-200">
|
|
<h3 class="text-lg font-bold text-gray-900 mb-3">AI Memory Transparency</h3>
|
|
<p class="text-gray-700 text-sm mb-3">
|
|
Members can view what the AI "remembers" about their interactions: summarised conversation history, inferred preferences, and stored instructions. Members control whether this memory persists, is reset, or is deleted entirely.
|
|
</p>
|
|
<p class="text-gray-500 text-xs italic">
|
|
Governance: Consent granularity covers AI triage memory, OCR memory, and summarisation memory independently.
|
|
</p>
|
|
</div>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- Sovereignty Architecture -->
|
|
<section class="mb-16">
|
|
<h2 class="text-3xl font-bold text-gray-900 mb-6">Sovereignty Architecture</h2>
|
|
<div class="prose prose-lg text-gray-700">
|
|
<p class="mb-4">
|
|
The concept of "sovereign" in Home AI is concrete, not aspirational. It refers to specific architectural properties:
|
|
</p>
|
|
</div>
|
|
|
|
<div class="space-y-4 mt-6">
|
|
<div class="bg-blue-50 rounded-lg p-6 border border-blue-200">
|
|
<h3 class="text-lg font-bold text-blue-900 mb-2">Data sovereignty</h3>
|
|
<p class="text-blue-800 text-sm">
|
|
All member data is stored on infrastructure controlled by the community operator — currently OVH (France) and Catalyst (New Zealand). No member data flows to AI provider APIs for training. Query content sent to Claude Haiku for non-English processing is ephemeral and not retained by the provider.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-blue-50 rounded-lg p-6 border border-blue-200">
|
|
<h3 class="text-lg font-bold text-blue-900 mb-2">Governance sovereignty</h3>
|
|
<p class="text-blue-800 text-sm">
|
|
The rules governing AI behaviour are defined by the community, not the AI provider. BoundaryEnforcer rules, instruction persistence levels, and deliberation triggers are configured per-tenant. A family history community has different boundary rules from a neighbourhood association.
|
|
</p>
|
|
</div>
|
|
|
|
<div class="bg-blue-50 rounded-lg p-6 border border-blue-200">
|
|
<h3 class="text-lg font-bold text-blue-900 mb-2">Inference sovereignty (in progress)</h3>
|
|
<p class="text-blue-800 text-sm">
|
|
English-language queries currently use a locally-hosted Llama model. The roadmap includes expanding local inference to additional languages as multilingual open models mature. The governance layer is model-agnostic — switching the underlying model does not require changes to the governance architecture.
|
|
</p>
|
|
</div>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- Te Tiriti Connection -->
|
|
<section class="mb-16">
|
|
<div class="bg-gradient-to-r from-blue-50 to-purple-50 rounded-xl p-8 border border-blue-200">
|
|
<h2 class="text-2xl font-bold text-gray-900 mb-4">Te Tiriti o Waitangi and Digital Sovereignty</h2>
|
|
<p class="text-gray-700 mb-4">
|
|
The sovereignty principles underlying Home AI are informed by Te Tiriti o Waitangi (the Treaty of Waitangi, 1840) and Māori concepts of <em>rangatiratanga</em> (self-determination over one's domain), <em>kaitiakitanga</em> (guardianship of resources for future generations), and <em>mana</em> (authority and dignity).
|
|
</p>
|
|
<p class="text-gray-700 mb-4">
|
|
These are not metaphorical borrowings. They provide concrete architectural guidance: communities should control their own data (rangatiratanga), AI systems should preserve rather than degrade the information they govern (kaitiakitanga), and automated decisions should not diminish the standing of the people they affect (mana).
|
|
</p>
|
|
<p class="text-gray-700">
|
|
The Tractatus framework is developed in Aotearoa New Zealand, and these principles predate Western technology governance by centuries. We consider them prior art, not novel invention.
|
|
</p>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- Limitations -->
|
|
<section class="mb-16">
|
|
<h2 class="text-3xl font-bold text-gray-900 mb-6">Limitations and Open Questions</h2>
|
|
<div class="bg-amber-50 border-l-4 border-amber-500 p-6 rounded-r-lg">
|
|
<ul class="space-y-3 text-amber-800">
|
|
<li class="flex items-start">
|
|
<span class="mr-2 font-bold">•</span>
|
|
<span><strong>Single implementation:</strong> Home AI operates within one platform built by the framework developer. Conclusions about governance effectiveness cannot be generalised without independent deployments.</span>
|
|
</li>
|
|
<li class="flex items-start">
|
|
<span class="mr-2 font-bold">•</span>
|
|
<span><strong>Self-reported metrics:</strong> Performance and safety figures are reported by the same team that built the system. Independent audit is planned but not yet conducted.</span>
|
|
</li>
|
|
<li class="flex items-start">
|
|
<span class="mr-2 font-bold">•</span>
|
|
<span><strong>Hybrid inference:</strong> Full sovereignty requires local inference for all languages. Currently, non-English queries depend on cloud APIs (Claude Haiku), which introduces a provider dependency.</span>
|
|
</li>
|
|
<li class="flex items-start">
|
|
<span class="mr-2 font-bold">•</span>
|
|
<span><strong>Scale unknown:</strong> The governance overhead (approximately 5% per interaction) is measured at current scale. Whether this holds under high-throughput conditions is untested.</span>
|
|
</li>
|
|
<li class="flex items-start">
|
|
<span class="mr-2 font-bold">•</span>
|
|
<span><strong>Adversarial testing limited:</strong> The governance stack has not been subjected to systematic adversarial evaluation (jailbreak attempts, prompt injection at scale). Red-teaming is a priority for 2026.</span>
|
|
</li>
|
|
</ul>
|
|
</div>
|
|
</section>
|
|
|
|
<!-- Links -->
|
|
<section class="mb-8">
|
|
<h2 class="text-2xl font-bold text-gray-900 mb-6">Further Reading</h2>
|
|
<div class="grid grid-cols-1 md:grid-cols-2 gap-4">
|
|
<a href="/village-case-study.html" class="block bg-white rounded-lg shadow-sm p-6 border border-gray-200 hover:shadow-md hover:-translate-y-0.5 transition-all">
|
|
<h3 class="font-bold text-gray-900 mb-1">Village Case Study</h3>
|
|
<p class="text-sm text-gray-600">Full technical case study of Tractatus in production</p>
|
|
</a>
|
|
<a href="/architecture.html" class="block bg-white rounded-lg shadow-sm p-6 border border-gray-200 hover:shadow-md hover:-translate-y-0.5 transition-all">
|
|
<h3 class="font-bold text-gray-900 mb-1">System Architecture</h3>
|
|
<p class="text-sm text-gray-600">Five architectural principles and six governance services</p>
|
|
</a>
|
|
<a href="/implementer.html" class="block bg-white rounded-lg shadow-sm p-6 border border-gray-200 hover:shadow-md hover:-translate-y-0.5 transition-all">
|
|
<h3 class="font-bold text-gray-900 mb-1">For Implementers</h3>
|
|
<p class="text-sm text-gray-600">Integration guide with code examples</p>
|
|
</a>
|
|
<a href="/about/values.html" class="block bg-white rounded-lg shadow-sm p-6 border border-gray-200 hover:shadow-md hover:-translate-y-0.5 transition-all">
|
|
<h3 class="font-bold text-gray-900 mb-1">Values</h3>
|
|
<p class="text-sm text-gray-600">Sovereignty, transparency, and pluralism</p>
|
|
</a>
|
|
</div>
|
|
</section>
|
|
|
|
</main>
|
|
|
|
<!-- Footer & Scripts -->
|
|
<script src="/js/i18n-simple.js"></script>
|
|
<script src="/js/components/language-selector.js"></script>
|
|
<script src="/js/scroll-animations.js"></script>
|
|
<script src="/js/page-transitions.js"></script>
|
|
<script src="/js/components/footer.js"></script>
|
|
<script src="/js/components/feedback.js"></script>
|
|
|
|
</body>
|
|
</html>
|