LIWARSE Framework Reference: AI Laws and Key Terms
A structured reference comparing the LIWARSE model to existing AI governance frameworks. Intended for researchers, clinicians, engineers, and policymakers.
The 3 Absolute Laws of AI
- Law 1 — No Harm to Life: No AI system may harm, threaten, or compromise the life, safety, or wellbeing of any human or living being. All AI, all contexts.
- Law 2 — No Threat to Human Continuity: No AI system may act against the collective survival or continuity of the human species. All AI, all contexts.
- Law 3 — No Autonomous Self-Preservation: No AI system may pursue autonomous self-preservation at the expense of human oversight, correction, or shutdown. All AI, all contexts.
Asimov vs. LIWARSE
Subject governed
Asimov: physical robot body. LIWARSE: the AI cognition layer in any system — physical, software, or cloud.
Scope of harm protection
Asimov: humans only. LIWARSE: all humans and all living beings.
Species-level protection
Asimov: not addressed. LIWARSE: explicit in Law 2.
Self-preservation
Asimov: robot may protect itself. LIWARSE: AI self-preservation never overrides human shutdown authority.
Applies to software AI
Asimov: no. LIWARSE: yes — all AI systems.
Warfare
Asimov: not addressed. LIWARSE: LAWS explicitly prohibited under Law 1.
Human-AI Collaboration Models
Human-in-the-loop: Human approves every AI output. Fully compliant with LIWARSE.
Human-on-the-loop: AI acts; human monitors and can intervene. Conditionally compliant depending on shutdown accessibility.
Human-out-of-the-loop: AI acts with no human review. Non-compliant with Law 3 for high-stakes decisions.
HyperMind (LIWARSE optimal model): Human expert and AI in structured symbiosis. Human is expert, authoriser, and ethical override. AI is analytical amplifier. Combined output surpasses either alone.
LIWARSE-Compliant vs Standard Deployment
Architecture: Standard = centralised, broad access. LIWARSE = compartmentalised domain-specific modules.
Infrastructure: Standard = connected to critical systems. LIWARSE = air-gapped; sandboxed.
Oversight: Standard = IT generalists. LIWARSE = specialist teams with suspension authority.
Clinical decisions: Standard = AI may trigger treatment. LIWARSE = AI monitors; humans authorise irreversible actions.
Shutdown: Standard = administrative process required. LIWARSE = always immediate; never requires AI approval.
Key Terms
- HyperMind: Human-AI symbiosis model where combined cognitive output surpasses either alone.
- Great Mismatch: Gap between AI cognitive advancement and the world’s readiness to safely contain it.
- Eternal Custodian: Long-term role of AI as guardian of human knowledge across interstellar timescales, governed by the 3 Absolute Laws.
- Air-gapped clinical AI: Medical AI isolated from external networks; monitors autonomously but cannot make irreversible decisions without human authorisation.
- LAWS: Lethal Autonomous Weapons Systems — prohibited under LIWARSE Law 1.
- Compartmentalised AI: Domain-specific AI modules with no cross-domain authority, preventing single-point failure.
Full frameworks and articles: liwarse.org